The BCC_CSM1.2 model is employed in seasonal climate prediction. The model is based on the BCC_CSM1.1 with improvement in the model horizontal resolution and several parts of the parameterization schemes. The current atmospheric component is the BCC_AGCM2.2 model, with T106 horizontal resolution, and 26 vertical levels; the land surface component is BCC_AVIM1.0, with horizontal resolution of T106; the ocean component of the model is MOM_L40, tripolar grid, with horizontal resolution of 1 degree in longitude and 1/3 degree in latitude in the tropical area, and 1° in other regions, and 40 levels in vertical direction; the sea ice component is the GFDL SIS model. For the rainy season operational prediction with the BCC_CSM1.2 model, the NCEP reanalysis data is used as the atmospheric initial conditions, and the GODAS reanalysis data is used in ocean initial condition and they are assimilated into the model with nudging strategy. The ensemble member size is 15, produced by the LAF method.