diff --git a/soh/src/overlays/actors/ovl_En_Ru2/z_en_ru2.c b/soh/src/overlays/actors/ovl_En_Ru2/z_en_ru2.c
index ea53206f8..433c90521 100644
--- a/soh/src/overlays/actors/ovl_En_Ru2/z_en_ru2.c
+++ b/soh/src/overlays/actors/ovl_En_Ru2/z_en_ru2.c
@@ -820,6 +820,19 @@ void func_80AF3F20(EnRu2* this, GlobalContext* globalCtx) {
 void EnRu2_Draw(Actor* thisx, GlobalContext* globalCtx) {
     EnRu2* this = (EnRu2*)thisx;
 
+    // FAST3D: This is a hack for the issue of both TEXEL0 and TEXEL1 using the same texture with different settings.
+    // Ruto's earring uses both TEXEL0 and TEXEL1 to render. The issue is that it never loads anything into TEXEL1, so
+    // it reuses whatever happens to be there, which is the water temple brick texture. It just so happens that the
+    // earring texture loads into the same place in tmem as the brick texture, so when it comes to rendering, TEXEL1
+    // uses the earring texture with diffrent clamp settings, and it displays without noticeable error. However, both
+    // texel samplers are not intended to be used for the same texture with different settings, so this misuse confuses
+    // our texture cache, and we load the wrong settings for the earrings texture. This patch is a hack that replaces
+    // TEXEL1 with TEXEL0, which is most likely the original intention, and all is well.
+    Gfx* gfx = ResourceMgr_LoadGfxByName(gAdultRutoHeadDL);
+    Gfx patch = gsDPSetCombineLERP(TEXEL0, 0, PRIMITIVE, 0, TEXEL0, 0, ENVIRONMENT, 0, 0, 0, 0, COMBINED, TEXEL0, 0,
+                                  PRIM_LOD_FRAC, COMBINED);
+    gfx[0xA2] = patch;
+
     if ((this->drawConfig < 0) || (this->drawConfig >= ARRAY_COUNT(sDrawFuncs)) ||
         (sDrawFuncs[this->drawConfig] == 0)) {
         // "Draw Mode is improper!"