Soil bulk density is vital for national carbon inventory and hydrological modeling study. In this study 19 intact paddy soil cores at a depth of 0–1 m located in Jinjing Town of Hunan Province in subtropical region were collected. By utilizing soil organic carbon content, texture and soil depth as prediction parameters, a regressive model for estimating soil bulk density was developed and compared to existing models. Results indicated that: i) soil bulk density is significantly influenced by soil organic carbon content with r = –0.81, followed by soil depth and texture, but influenced less by soil clay content. Loam and silt loamy were accounted for most of studied soil textures. The bulk density ranged from 0.79 to 1.66 g/cm3 with an average of 1.27 g/cm3 for topsoil, but varied between 1.32 and 1.80 g/cm3 with an average of 1.58 g/cm3 for subsoil, and ii) the adjusted determined coefficient of derived model was 0.75 with 0.01 of ME and 0.10 of RMSE, respectively. This study also found that the Kaur (2002) and Manrique (1991) models could make good predictions of soil bulk density, but still not as good as the model developed in this study. Thus, the soil bulk density prediction model developed in this study will play a crucial role in estimating carbon stock of paddy fields as well as in applying hydrological models in the region.