It is of agronomic importance to apply nitrogen (N), but it has high environmental risks in reclaimed saline soils. Therefore, we should apply N fertilizer at an appropriate rate to increase crop yield but decrease N losses. In this soil column experiment, rice yield, N uptake, and ammonia (NH3) and nitrous oxide (N2O) losses were measured in four treatments with no N application (control) and with N applications of 160, 200, and 240 kg/ha (N160, N200, and N240, respectively). The results show that grain yield, spike number, and thousand-kernel weight increased with increases in N application rate, but there was no significant difference in grain yield between N200 and N240. However, the kernels per spike increased first and then decreased with the increase in N application, of which N200 was recorded to have the highest kernels per spike value, which was 16.8 and 9.8% higher than those of N160 and N240, respectively. Total NH3 volatilization of the rice season increased with increasing N input, especially during the first and second supplementary fertilization stages. The NH4+-N concentration of overlying water was relatively lower under the N200 treatment in these two stages, and the yield-scaled NH3 volatilization and the emission factor were the lowest in N200, which were 26.2–27.8% and 4.0–21.0% lower than those of N160 and N240, respectively. Among the three N-applied treatments, N2O losses and the emission factor as well as the yield-scaled N2O emissions were the lowest under the N200 treatment, which had 34.7% and 78.9% lower N2O emissions and 57.8% and 83.5% lower emission factors than those of the N160 and N240 treatments, respectively. Moreover, the gene copies of AOA and AOB amoA, nirS, and nirK in cultivated layer soils all reached the minimum under the N200 treatment. According to the comprehensive effects of N fertilizer on rice grain yield and NH3 and N2O losses, we recommend applying 200 kg/ha to reclaimed saline soil to ensure crop yield and reduce N losses.