In this study, we evaluate the global sea ice modeling capability of the Beijing Climate Center Climate System Model (BCC_CSM). Comparative analysis results indicate that the model simulates the spatial and temporal distribution characteristics of global sea ice extent and thickness well and that its simulation performance in the Southern Hemisphere is better than that in the Northern Hemisphere. The maximum simulation bias of the annual mean sea ice extent (SIE) occurs in the Sea of Okhotsk, the Bering Sea, and the Barents Sea. The spatial distribution results for simulated sea ice thickness are similar to those from observation data, with thinner sea ice in the Northern Hemisphere winter. The mean annual cycle of sea ice extent has a negative bias in summer and a positive bias in winter, as compared with observation data. The annual mean SIE in the past 60 years in both the Northern and Southern Hemispheres is excessive, while the positive bias in the Southern Hemisphere is smaller. However, the simulated interannual variation trend in the Northern Hemisphere is more accurate. In addition, lower net radiation results lead to anomalous cold sea surface temperatures, which may be the main reason for the sea ice simulation error.