diff --git a/scripts/addons_core/io_scene_fbx/import_fbx.py b/scripts/addons_core/io_scene_fbx/import_fbx.py
index 13daf585cba..0c95352605e 100644
--- a/scripts/addons_core/io_scene_fbx/import_fbx.py
+++ b/scripts/addons_core/io_scene_fbx/import_fbx.py
@@ -1784,6 +1784,11 @@ def blen_read_geom_layer_normal(fbx_obj, mesh, xform=None):
     return False
 
 
+def normalize_vecs(vectors):
+    norms = np.linalg.norm(vectors, axis=1, keepdims=True)
+    np.divide(vectors, norms, out=vectors, where=norms != 0)
+
+
 def blen_read_geom(fbx_tmpl, fbx_obj, settings):
     # Vertices are in object space, but we are post-multiplying all transforms with the inverse of the
     # global matrix, so we need to apply the global matrix to the vertices to get the correct result.
@@ -1914,6 +1919,10 @@ def blen_read_geom(fbx_tmpl, fbx_obj, settings):
         clnors = np.empty(len(mesh.loops) * 3, dtype=bl_nors_dtype)
         mesh.attributes["temp_custom_normals"].data.foreach_get("vector", clnors)
 
+        clnors = clnors.reshape(len(mesh.loops), 3)
+        normalize_vecs(clnors)
+        clnors = clnors.reshape(len(mesh.loops) * 3)
+
         # Iterating clnors into a nested tuple first is faster than passing clnors.reshape(-1, 3) directly into
         # normals_split_custom_set. We use clnors.data since it is a memoryview, which is faster to iterate than clnors.
         mesh.normals_split_custom_set(tuple(zip(*(iter(clnors.data),) * 3)))