As　human　activities　highly　depend　on　the　land　resources　and　changed　the　land　cover　(LC)　condition,　the　relationship　between　LC　and　nighttime　light　(NTL)　intensity　has　been　widely　analyzed　to　support　the　foundation　of　NTL　applications　and　help　explain　the　drivers　of　urban　economic　development.　However,　previous　studies　always　paid　attention　to　the　effect　of　each　LC　type　on　NTL　intensity,　with　limited　consideration　of　the　joint　effects　of　any　two　LC　types.　To　fill　this　gap,　this　study　measured　the　land　cover　spatial　combination　(LCSC)　by　using　a　spatial　adjacency　matrix,　and　then　analyzed　its　impacts　on　NTL　intensity　based　on　an　extreme　gradient　boosting　(XGBoost)　regression　model　with　the　assistant　of　sharpley　additive　explanations　(SHAP)　method.　Our　results　presented　that　the　LCSC　can　better　(R2　of　82.4%　and　98.1%　in　2010　and　2020)　explain　the　relationship　between　LC　and　NTL　intensity　with　the　traditional　LC　metrics　(e.g.,　area　and　patch　count),　since　the　LCSC　is　much　more　sensitive　to　the　diverse　land　functions.　It　is　noteworthy　that　the　impacts,　as　well　as　their　dynamics,　of　LCSC　between　any　two　LC　types　on　NTL　intensity　are　various.　LCSC　associated　with　artificial　surface　contributed　more　to　NTL　intensity.　In　detail,　the　LCSC　of　water/wetland　and　artificial　surface　can　increasingly　promote　the　NTL　intensity　while　the　LCSC　of　grassland/forest　and　artificial　surface　has　a　decreasing　or　inverse　U-shaped　contribution　to　NTL　intensity.　Whereas　LCSC　associated　with　non-artificial　surface　were　not　conducive　to　the　increase　in　NTL　intensity　due　to　high　vegetation　density.　We　also　provided　three　implications　to　help　further　urbanization　process　and　discussed　the　applications　of　LCSC.　©　2023,　The　Author(s).