diff --git a/src_c/alphablit.c b/src_c/alphablit.c
index e975218fda..a4867053b7 100644
--- a/src_c/alphablit.c
+++ b/src_c/alphablit.c
@@ -2841,7 +2841,8 @@ premul_surf_color_by_alpha(SDL_Surface *src, SDL_Surface *dst)
         // since we know dst is a copy of src we can simplify the normal checks
 #if !defined(__EMSCRIPTEN__)
 #if SDL_BYTEORDER == SDL_LIL_ENDIAN
-    if ((PG_SURF_BytesPerPixel(src) == 4) && pg_has_avx2()) {
+    if ((PG_SURF_BytesPerPixel(src) == 4) &&
+        (src->pitch % PG_SURF_BytesPerPixel(src) == 0) && pg_has_avx2()) {
         premul_surf_color_by_alpha_avx2(src, dst);
         return 0;
     }
@@ -2873,6 +2874,8 @@ premul_surf_color_by_alpha_non_simd(SDL_Surface *src, SDL_Surface *dst)
     int srcbpp = PG_FORMAT_BytesPerPixel(srcfmt);
     int dstbpp = PG_FORMAT_BytesPerPixel(dstfmt);
     Uint8 *src_pixels = (Uint8 *)src->pixels;
+    int srcskip = src->pitch - (width * srcbpp);
+    int dstskip = dst->pitch - (width * dstbpp);
     Uint8 *dst_pixels = (Uint8 *)dst->pixels;
 
     int srcpxskip = PG_SURF_BytesPerPixel(src);
@@ -2898,6 +2901,8 @@ premul_surf_color_by_alpha_non_simd(SDL_Surface *src, SDL_Surface *dst)
                 dst_pixels += dstpxskip;
             },
             n, width);
+        src_pixels += srcskip;
+        dst_pixels += dstskip;
     }
 }
 
diff --git a/src_c/simd_blitters_sse2.c b/src_c/simd_blitters_sse2.c
index 65bb926721..257f48de96 100644
--- a/src_c/simd_blitters_sse2.c
+++ b/src_c/simd_blitters_sse2.c
@@ -793,7 +793,9 @@ premul_surf_color_by_alpha_sse2(SDL_Surface *src, SDL_Surface *dst)
     int width = src->w;
     int height = src->h;
     Uint32 *srcp = (Uint32 *)src->pixels;
+    int srcskip = src->pitch - width * PG_SURF_BytesPerPixel(src);
     Uint32 *dstp = (Uint32 *)dst->pixels;
+    int dstskip = dst->pitch - width * PG_SURF_BytesPerPixel(dst);
 
     SDL_PixelFormat *srcfmt = src->format;
     Uint32 amask = srcfmt->Amask;
@@ -846,6 +848,8 @@ premul_surf_color_by_alpha_sse2(SDL_Surface *src, SDL_Surface *dst)
                 ++dstp;
             },
             n, width);
+        srcp = (Uint32 *)((Uint8 *)srcp + srcskip);
+        dstp = (Uint32 *)((Uint8 *)dstp + dstskip);
     }
 }
 
diff --git a/test/surface_test.py b/test/surface_test.py
index ca60ccb8e0..df64e6882d 100644
--- a/test/surface_test.py
+++ b/test/surface_test.py
@@ -4038,6 +4038,26 @@ def test_surface_premul_alpha(self):
                         ),
                     )
 
+        def create_surface_from_byte_width(byte_width):
+            surf_height = 5
+            byte_data = bytes(byte_width * surf_height)  # 50 bytes
+            surf_width = byte_width // 4  # width = 2
+
+            dest = pygame.image.frombuffer(
+                byte_data, (surf_width, surf_height), "RGBA", pitch=byte_width
+            )
+            dest.fill((120, 50, 70, 200))
+            return dest
+
+        test_surf = create_surface_from_byte_width(10)
+        test_surf = test_surf.premul_alpha()
+
+        for y in range(0, test_surf.get_height()):
+            for x in range(0, test_surf.get_width()):
+                self.assertEqual(
+                    test_surf.get_at((x, y)), pygame.Color(94, 39, 55, 200)
+                )
+
     def test_surface_premul_alpha_ip(self):
         """Ensure that .premul_alpha_ip() works correctly"""