This study explored the effects of the sowing stage and nitrogen application rate on the grain yield and its allocation of light-temperature resources over a 9-year experiment from 2011 to 2019. Measurement indicators include the effective accumulative temperature on different growth durations, leaf area index (LAI), above-ground biomass production, and harvest index (HI). Methods: A split-plot design was arranged in the treatment, with N supply as the main plot and the sowing stage as the subplot. The main plots consisted of two nitrogen treatments: low nitrogen (LN: 120 kg ha−1) and high nitrogen (HN: 180 kg ha−1). The subplots contained two sowing stages: the early sowing stage (ES) and the late sowing stage (LS). Results: Compared with LNLS, LNES, and HNLS from 2011 to 2019, HNES of HHZ increased the grain yield by 9.5%, 2.5%, and 5.3%, while the difference in grain yield in YY8 was higher than HHZ, especially under HNES. Compared with LNLS, LNES, and HNLS from 2011 to 2019, HNES of HHZ increased the panicle number by 6.0%, 5.9%, and 1.0%, and HNES of YY8 increased by 12.7%, 11.4%, and 3.8%. Compared with HNLS of HHZ, LNES, LNLS, and HNES decreased the spikelets per panicle by 2.3%, 2.9%, and 1.1%, and decreased by 3.5%, 1.9%, and 2.2% in YY8. The early sowing or increasing N supply significantly increased the dry matter accumulated, grain weight, LAI, and HI. The higher grain yield in LNES was more closely related to the average temperature and the number of spikelets per panicle. The grain yield in HNES was more dependent on the effective accumulative temperature. Conclusions: Sowing in mid-May and increasing the N application (180 kg ha−1) are beneficial to the allocation of light temperature and the increase in yield. Therefore, this research provides a theoretical basis for improving rice yield and optimizing the utilization of light-temperature resources in the future.