Increased plant density with low N rate was a recommended strategy to increase grain yield and N use efficiency (NUE); however, grain yield, NUE and the total N uptake (TNU) responses of hybrid rice to this strategy at different yield levels (medium yielding site (MYS) Luzhou City and high yielding site (HYS) Deyang City had not been described. Field experiments with hybrid rice Rongyou1015 were conducted to study the effects of two plant densities. High plant density (HD), low plant density ( LD) and four N rates (without N, N0; a recommended N rate of 195 kg ? hm-2, NCK; a 23％ reduction in N rate, N-23％; a 46％ reduction in N rate, N-46％) on yield attributes, grain yield, TNU and NUE of hybrid rice were studied under different yield levels in 2016-2017. The results showed that the grain yield and NUE of hybrid rice in response to plant density and N rate varied with yield levels. For MYS, reducing N rate by 46％ result in significantly lower grain yield at LD treatment; whereas at HD treatment the grain yield of hybrid rice under N-46％ and NCK were equal. For HYS, reducing N rate by 46％ result in significantly lower grain yield regardless of low plant density and high plant density; however, a reduction in N rate by 23％ increased grain yield, AEN by 36％, PFPN by 31％ and REN by 11％ over NCK at HD treatment. Higher grain yield of hybrid rice under the combination of HD with low N rate was attributable to improvement in spikelets per panicle and harvest index. The results suggested that high plant density with low N rate might be an effective approach to improve grain yield and NUE in rice production, but reduction in N application rate was determined, according to yield levels.