In　designing　highly　efficient　CO2　reduction　reaction　(CRR)　photocatalysts　with　excellent　selectivity　and　efficiency,　a　key　limitation　is　the　poor　understanding　on　the　mechanism　response　of　the　active　sites　of　catalysts　to　CRR　selectivity　and　activity.　Herein,　it　is　revealed　how　the　concentration　of　point　defect　affects　the　CRR　selectivity　and　activity　of　catalysts.　Quasi-2D　NiO　nanosheets　(NSs)　that　are　composed　of　NiO　nanoparticles　(NPs),　and　finely　tuned　the　oxygen　vacancies　(OVs)　concentration　of　the　NSs　by　regulating　the　grain　size　(approximate　to　5-25　nm)　of　NPs　are　created.　The　NiO　with　moderate　OVs　concentration　has　the　highest　photocatalytic　CRR　efficiency　and　selectivity　(V-CO　=　17.2　mu　mol　h(-1),　98.3%),　outperforming　other　prepared　NiO　catalysts　and　reported　Ni-based　photocatalysts.　Density　functional　theory　calculation　associated　with　the　CO2　temperature-programmed　desorption　confirms　that　the　moderate　OVs　concentration　enables　strong　CO2　binding　to　promote　CO2　adsorption　and　activation　and　allows　efficient　charge　transfer.　In　contrast,　excessive　OVs　reduce　the　CO2　binding　affinity　and　restrain　charge　mobility,　both　detrimental　to　the　CRR　performance.