A two-year field experiment was conducted in the Taihu Lake Region with five treatments:no nitrogen (N) input (CK), conventional N broadcasting (CN), reduced N broadcasting (RN), reduced N with side deep fertilization (RNS), and reduced N with point deep fertilization (RNP), to explore the effects of different deep fertilization methods on ammonia volatilization (NH₃) and N use efficiency (NUE) in paddy fields. The results showed that compared to surface application treatments (CN and RN), RNS and RNP reduced ammonia volatilization by 30.95%-41.54% and 66.71%-72.23% (P<0.05), respectively, owing to the reduced NH₄⁺-N concentration and pH in floodwater. RNP significantly promoted the root growth of rice, and increased soil available N content in root zone than RN, thereby increasing rice yield (6.23%) and NUE (50.15%), and decreasing soil N surplus (63.92%) (P<0.05). RNS markedly reduced soil N surplus (29.20%) (P<0.05), but N uptake and NUE of rice were not significantly increased versus CN. In contrast with RNS, RNP further decreased NH₃ volatilization (50.84%) and soil N surplus (51.07%), and increased NUE (40.40%) (P<0.05). Overall, RNP can achieve the maximum agronomic and environmental benefits, but it is difficult to be widely used in agriculture production owing to the limitation of the development of point-deep fertilization machines and fertilizer granulation technology, whereas side-deep fertilization method is feasible to be adopted in large-scale intensive production of rice due to the higher agronomic and environmental benefits.