Abstract: 5 types of controlled release fertilizer (LP) were chosen and basal applied with various rates on early rice and late rice in red soil paddy field of Qiyang, Hunan province of Southern China. The effects of controlled-release fertilizer on rice yield and N dynamics in surface soil and water were studied. Results showed that the NH4+-N concentration in surface water of LP plots were much lower than that of urea within first 7-10 days. It was only about one tenth of that of urea. NH4+-N content in soil of cultivated layer of LP plots was also much lower than that of urea within first ten days. It was around one tenth as well. Therefore, it could deduce N losses from runoff and ammonia volatilization from LP plots should be much lower than that of urea.For early rice, application N 75kg/ha of LPS40 or LPS60, the yield was nearly same as that of urea with application rate being N 150kg/ha; but then, increased application rate of LPS40 or LPS60 to 150kg/ha could not increase yield further under conventional transplanting density. This indicated that for early rice, both LPS40 and LPS60 were suitable and the rational application rate was N 75 kg/ha. However, because N release from LPS40 and LPS60 was longer than that of urea, the yield of ratooning rice applied LPS40 and LPS60 with N 150kg/ha rate and 1.5 times higher transplanting density increased by 10% and 16%, respectively. For late rice, band application using LP70 (50%) + LPS80 (50%) could improve rice development. Broadcast and band application N 75kg/ha, rice yield increased by 18% and 25% respectively comparing with urea N 150 (kg/ha.) While yield applied LP N 150kg/ha was not different significantly with that of urea. Thus, the recommend rate of N is (75 kg/ha) rate with band application for late rice production.