Biochar application in rice production reduces nitrogen loss and greenhouse gases. We conducted in situ experiments for 3 years, with N210B0 (210 kg N ha-1) as the control. Two biochar application methods (B1:15 t ha-1 biochar applied once and B2: biochar applied three times at 5 t ha-1 yr-1) combined with two nitrogen levels (N210: 210 kg N ha-1 and N168: 168 kg N ha-1) were used. Soil physicochemical properties, CH4 and N2O emissions, functional gene abundance, rice yield, and nitrogen use efficiency were analyzed. Both methods improved the physicochemical properties of the soil, however, B1 was less effective than B2 in increasing soil pH, bulk density, organic carbon, total nitrogen, and microbial biomass nitrogen in year 3. B1 had a higher CH4 emission mitigation effect than B2 in 3 consecutive years, mainly due to the higher pmoA gene abundance. B1 showed a higher reduction effect of N2O emissions compared to B2 in year 1, but the opposite was observed in years 2 and 3. B2 had a higher abundance of AOB, nirK, and nosZ genes compared to B1 in year 3. Compared with N210B0, rice yields were increased by 9.1 %, 9.6 %, and 3.6 % with N210B1, N210B2, and N168B2, respectively, over 3 years, while N168B1 improved yields in the previous 2 years. Biochar improved nitrogen use efficiency over 3 consecutive years directly due to increased use efficiency of panicle fertilizer; the effect of B1 was greater than that of B2 during years 1 and 2, while the opposite was observed in year 3. Both Biochar applied once and three times appeared to be promising practices to increase yield and mitigate GHGs. From the GHGI perspective, the biochar applied once combined with 168 kg N ha-1 can further improve nitrogen use efficiency, and reduce GHGs without hindering improvements in rice yield.