Abstract: In this paper, the authors assess the summer prediction skill of retrospective forecasts of global midlatitudes and low latitudes sea surface temperatures (SSTs) in the Beijing Climate Center Climate System Model (BCC_CSM).Results indicate that the SST forecasts exhibit certain skills in middle and low latitude areas and the skills over much of the low latitude areas are better.Further analysis reveals that model forecasts in different areas of the sea have different skills.Specifically, the SST forecasts exhibit significant skills over much of the North Pacific and the Niño 3.4 index in summer, followed by the Indian Ocean, Atlantic Ocean, and the Indian Ocean Basin-wide Warming (IOBW), while the North Atlantic Tripole (NAT) shows a lower forecast skill.Recent investigations have revealed that prediction skill is closely related to the El Niño Southern Oscillation (ENSO) phases of the previous winter.When the previous winter phase is characterized by ENSO, SST forecasts exhibit higher skills than when preceded by a neutral phase, and the NAT index also has higher forecast skills.The ENSO phases of the previous winter have little effect on the skills of the Niño 3.4 and IOBW indexes.In most cases, as the lead time becomes longer, the model will show a drop in forecast skills for summer SSTs and these indexes.