Abstract: Marine heatwaves (MHWs), which have become increasingly frequent extreme climate events in recent years, exert significant impacts on both the ecological environment and the social economy. Consequently, the prediction of MHWs has garnered considerable attention. In this study, we utilize ocean heat content (OHC) to assess the short-term prediction skill of the First Institute of Oceanography-Climate Prediction System version 2.0 (FIO-CPS v2.0) for upper MHWs. Furthermore, we employ heat budget analysis to delve into the intricacies of MHW prediction. Our findings reveal that the prediction skill of FIO-CPS v2.0 for upper MHWs exhibits regional and seasonal variations. Specifically, the prediction skill is relatively higher in the tropical eastern Pacific, tropical western Pacific, and northeast Pacific. Notably, the prediction skill in the tropical eastern Pacific encounters a spring prediction barrier, whereas it performs better in the fall for the tropical western Pacific and the northeast Pacific. Upon further analysis, we discover that the upper MHW prediction skill of FIO-CPS v2.0 is closely linked to its underlying driving mechanisms. In the tropical eastern Pacific, the prediction skill stems from the El Niño-Southern Oscillation (ENSO) prediction skill. In the northeast Pacific, the surface and subsurface prediction skills are attributed to surface heat flux and horizontal advection, respectively. Meanwhile, in the tropical western Pacific, the prediction skill is provided by vertical advection. Additionally, FIO-CPS v2.0 demonstrates the ability to predict the changing characteristics of MHWs in the context of global warming. This study not only enhances our understanding of the prediction skill of the short-term climate prediction system FIO-CPS v2.0 but also offers a valuable reference for the short-term prediction of upper MHWs.