Legume-cereal inter-cropping can improve nitrogen (N) use efficiency and crop yields. Numerous attempts have been made to examine the effects of legume-cereal intercropping patterns in upland and dry fields. However, the rice-legume inter-cropping in irrigated farming systems remains to be investigated. In this study, a field experiment was conducted to evaluate the effects of rice-water mimosa (Neptunia oleracea) inter-cropping patterns on crop yield, N uptake, and soil N content dynamics at an irrigated farm located in Southern China. Four cropping patterns, namely, (1) rice monocropping with conventional N input (CK, N was applied at the level of 180 kg·hm^-2); (2) rice monocropping with low N input (LRM, N was applied at the level of 140 kg·hm^-2); (3) water mimosa monocropping (NOM, N was applied at the level of 140 kg·hm^-2); and (4) rice-water mimosa inter-cropping (LRN, N was applied at the level of 140 kg·hm^-2) were used for late rice in 2014 and early rice in 2015. Results showed that tiller number, leaf area index and above-ground biomass were significantly increased for the LRN pattern as compared to those for the CK and LRM patterns. Among the cropping patterns, tiller number, leaf area index, above-ground biomass and grain yield of rice at the edge rows (adjacent to water mimosa) were significantly higher than those at the inner rows. The rice yield was in the following order: LRN > CK > LRM. The yield of mimosa per unit of land area (based on the planted area) for the LRN pattern was lower than that for the NOM pattern. The land equivalent ratios calculated based on the NOM, LRN and CK patterns were 1.01 for late rice in 2014 and 1.2 for early rice in 2015, whereas the land equivalent ratios calculated based on the LRN, NOM, and LRM patterns were 1.12 for late rice in 2014 and 1.25 for early rice in 2015. The rice yield for the LRN pattern as compared to the CK increased 20.2% in 2014 and 35.3% in 2015, indicating that inter-cropping promoted crop yield per unit land area. The total N uptake amount of rice population was significantly lower for intercropping than for monocropping. After planting for two seasons, the soil total nitrogen (TN) content for the LRN pattern increased 0.21% as compared to that of CK pattern and increased 1.69% as compared to that of LRM pattern. These results suggested that rice-water mimosa inter-cropping can maintain a higher land equivalent ratio and enhance crop yield and soil TN content under lower N input.