diff --git a/B04_PreProcess/B04_PreProcess_Engine_Contour.py b/B04_PreProcess/B04_PreProcess_Engine_Contour.py
index 3ed70a52..7ede4004 100644
--- a/B04_PreProcess/B04_PreProcess_Engine_Contour.py
+++ b/B04_PreProcess/B04_PreProcess_Engine_Contour.py
@@ -10,7 +10,7 @@ from pathlib import Path
 from typing import Any
 
 import numpy as np
-from scipy.interpolate import RBFInterpolator, RectBivariateSpline
+from scipy.interpolate import RBFInterpolator
 from skimage import measure
 
 # 등고선 캐시 형식/추출 규칙이 바뀔 때 증가시킨다.
@@ -36,14 +36,17 @@ def extract_contours_from_grid(
 
     z_min = float(np.min(z_grid[finite_mask]))
     z_max = float(np.max(z_grid[finite_mask]))
+    span = z_max - z_min
+    if span <= 0:
+        return []
+    # 레벨 수 상한은 arange **앞에서** 건다. 뒤에 두면 표고 범위가 비정상적으로 넓을 때
+    # np.arange가 배열을 만들다 "Maximum allowed size exceeded"로 먼저 터진다.
+    if span / interval > 500:
+        interval = span / 100.0
     start_level = np.ceil(z_min / interval) * interval
     levels = np.arange(start_level, z_max, interval)
     if len(levels) == 0:
         return []
-    if len(levels) > 500:
-        new_interval = (z_max - z_min) / 100.0
-        levels = np.arange(np.ceil(z_min / new_interval) * new_interval, z_max, new_interval)
-        interval = new_interval
 
     contours_geojson_list: list[dict[str, Any]] = []
 
@@ -398,16 +401,15 @@ def extract_contours(
         )
 
     if representation == "bspline_surface":
+        # 모델 빌더와 같은 곡면을 다시 만든다 — 평활 계수·발산 방어를 공유한다.
+        from B04_PreProcess.B04_PreProcess_Engine_ModelContext import (
+            evaluate_nurbs_spline,
+            fit_nurbs_spline,
+        )
+
         control_x, control_y, control_z = data["control_x"], data["control_y"], data["control_z"]
         degree = int(data["degree"][0])
-        spline = RectBivariateSpline(
-            control_y,
-            control_x,
-            control_z,
-            kx=min(degree, len(control_y) - 1),
-            ky=min(degree, len(control_x) - 1),
-            s=float(len(control_x) * len(control_y)) * 0.01,
-        )
+        spline, z_range = fit_nurbs_spline(control_x, control_y, control_z, degree)
         x_coords, y_coords = _grid_axes(
             float(control_x[0]),
             float(control_x[-1]),
@@ -415,7 +417,9 @@ def extract_contours(
             float(control_y[-1]),
             target_grid_m,
         )
-        z_grid = np.asarray(spline(y_coords, x_coords), dtype=np.float32)
+        z_grid = evaluate_nurbs_spline(
+            spline, y_coords, x_coords, z_range, Path(model_npz_path).stem
+        ).astype(np.float32)
         valid_mask = _apply_footprint(
             model_npz_path, x_coords, y_coords, np.ones_like(z_grid, dtype=bool)
         )
diff --git a/B04_PreProcess/B04_PreProcess_Engine_ModelBuild.py b/B04_PreProcess/B04_PreProcess_Engine_ModelBuild.py
index 09f032c8..baa33c69 100644
--- a/B04_PreProcess/B04_PreProcess_Engine_ModelBuild.py
+++ b/B04_PreProcess/B04_PreProcess_Engine_ModelBuild.py
@@ -8,7 +8,7 @@ from pathlib import Path
 from typing import Any
 
 import numpy as np
-from scipy.interpolate import RBFInterpolator, RectBivariateSpline
+from scipy.interpolate import RBFInterpolator
 from scipy.spatial import Delaunay
 
 from B04_PreProcess.B04_PreProcess_Engine_ModelContext import (
@@ -17,6 +17,8 @@ from B04_PreProcess.B04_PreProcess_Engine_ModelContext import (
     artifact_size,
     atomic_npz,
     clip_and_compact_mesh,
+    evaluate_nurbs_spline,
+    fit_nurbs_spline,
     grid_faces,
     grid_vertices,
     with_footprint,
@@ -125,18 +127,13 @@ def build_nurbs(
     control_resolution = max(patch_size / max(controls - 1, 1), 0.25)
     x_control, y_control, z_control = context.grid(control_resolution)
     progress(30)
-    spline = RectBivariateSpline(
-        y_control,
-        x_control,
-        z_control,
-        kx=min(degree, len(y_control) - 1),
-        ky=min(degree, len(x_control) - 1),
-        s=float(len(x_control) * len(y_control)) * 0.01,
-    )
+    spline, z_range = fit_nurbs_spline(x_control, y_control, z_control, degree)
     x_preview, y_preview, _ = context.preview_grid(
         float(context.config["dtm_grid_resolution_meters"])
     )
-    z_preview = np.asarray(spline(y_preview, x_preview), dtype=np.float32)
+    z_preview = evaluate_nurbs_spline(spline, y_preview, x_preview, z_range, stem).astype(
+        np.float32
+    )
     progress(65)
     vertices = grid_vertices(x_preview, y_preview, z_preview)
     faces = grid_faces(len(y_preview), len(x_preview))
diff --git a/B04_PreProcess/B04_PreProcess_Engine_ModelContext.py b/B04_PreProcess/B04_PreProcess_Engine_ModelContext.py
index a113f652..e58fe9e4 100644
--- a/B04_PreProcess/B04_PreProcess_Engine_ModelContext.py
+++ b/B04_PreProcess/B04_PreProcess_Engine_ModelContext.py
@@ -7,6 +7,7 @@ GLB/PLY 프리뷰 및 npz 모델을 원자적으로 저장하는 공통 기능
 
 import hashlib
 import json
+import logging
 import math
 from dataclasses import dataclass, field
 from pathlib import Path
@@ -15,9 +16,12 @@ from typing import Any, Callable
 import numpy as np
 import trimesh
 from scipy import ndimage
+from scipy.interpolate import RectBivariateSpline
 
 from common_util.common_util_atomic import atomic_write_bytes, atomic_write_npz
 
+logger = logging.getLogger(__name__)
+
 MODEL_VERSION = 1
 MODEL_METHODS = ("tin", "dtm", "nurbs", "implicit", "meshfree")
 SOURCE_FILTER_LABELS = {"grid_min_z": "Grid Min-Z", "csf": "CSF", "pmf": "PMF"}
@@ -26,6 +30,62 @@ ProgressCallback = Callable[[int], None]
 # 대용량 포인트 배치 처리 크기
 _BATCH_SIZE = 500_000
 
+# NURBS 평활 계수 — 제어점당 허용 잔차제곱(m²).
+# 0.01(RMS 0.1m)은 요구가 너무 빡빡해 FITPACK이 제어점 수(122x113)보다 많은
+# knot(126x117)을 밀어넣고 계수가 발산했다 — csf 표고 1e105, grid_min_z 1e10
+# (2026-08-17 실측). 1.0(RMS 1m)이면 knot이 17~35로 떨어지고 세 지면 필터 모두
+# 데이터 표고 범위 안에 머문다. NURBS는 평활 곡면 표현이라 이 정도 완화가 맞다.
+NURBS_RESIDUAL_PER_CONTROL_POINT = 1.0
+
+
+def fit_nurbs_spline(
+    x_control: np.ndarray, y_control: np.ndarray, z_control: np.ndarray, degree: int
+) -> tuple[RectBivariateSpline, tuple[float, float]]:
+    """제어 격자에 B-spline 곡면을 맞추고, 발산 판정용 표고 허용 범위를 함께 준다.
+
+    모델 빌더와 등고선 엔진이 같은 곡면을 각각 만들므로 여기서 한 번에 정의한다.
+    """
+    spline = RectBivariateSpline(
+        y_control,
+        x_control,
+        z_control,
+        kx=min(degree, len(y_control) - 1),
+        ky=min(degree, len(x_control) - 1),
+        s=float(len(x_control) * len(y_control)) * NURBS_RESIDUAL_PER_CONTROL_POINT,
+    )
+    z_min = float(np.min(z_control))
+    z_max = float(np.max(z_control))
+    margin = max((z_max - z_min) * 0.5, 1.0)
+    return spline, (z_min - margin, z_max + margin)
+
+
+def evaluate_nurbs_spline(
+    spline: RectBivariateSpline,
+    y_coords: np.ndarray,
+    x_coords: np.ndarray,
+    z_range: tuple[float, float],
+    label: str = "",
+) -> np.ndarray:
+    """스플라인을 평가하고 표고 허용 범위를 벗어난 값을 잘라낸다.
+
+    평활 스플라인은 제어점이 성긴 구석에서 발산할 수 있다. 그대로 두면 float32
+    캐스팅에서 inf가 되고 프리뷰 색상(nan)과 등고선 레벨 산출(`np.arange`가
+    "Maximum allowed size exceeded"로 실패)까지 번진다.
+    """
+    z_values = np.asarray(spline(y_coords, x_coords), dtype=np.float64)
+    low, high = z_range
+    outside = np.count_nonzero(~np.isfinite(z_values) | (z_values < low) | (z_values > high))
+    if outside:
+        logger.warning(
+            "NURBS 곡면이 표고 범위를 벗어나 잘라냈습니다: %s%d개 셀 (허용 %.1f~%.1fm)",
+            f"{label} " if label else "",
+            int(outside),
+            low,
+            high,
+        )
+    z_values = np.nan_to_num(z_values, nan=low, posinf=high, neginf=low)
+    return np.clip(z_values, low, high)
+
 
 def config_signature(config: dict[str, Any]) -> str:
     """지오메트리 원본과 무관한 등고선·스무딩 설정을 제외한 캐시 서명."""