As one of the prominent climatic component of tropical climate system, the annual cycle of the equatorial Atlantic cold tongue region has been a subject of significant research. In this work, we examine the fidelity of the key part of the annual cycles, namely, the upwelling annual cycles, from the simulations of the Coupled Model Intercomparison Project Phase 5 (CMIP5). Utilizing the simple framework we have recently developed based on the Zebiak-Cane (ZC) ocean model, we first tested the validity of the simple framework of upwelling as a combination of the local wind-driven Ekman upwelling and nonlocal wind-driven wave upwelling. It is demonstrated that the theoretically reconstructed upwelling annual cycles to large extent are in good agreement with the simulations in these coupled climatic models. Comparing with observations, we show that the semiannual component of the upwelling is systematically too strong. This significant bias is attributable to the dynamic amplification to the wind-driven wave upwelling originated from a less significant bias in the semiannual harmonic component in the central equatorial wind field. The latter is dynamically amplified in the wind-driven wave upwelling owing to its relatively fast time scale. Thus, the realistic simulations of equatorial annual cycles may sensitively depend on good simulations of the semiannual harmonic components in equatorial wind fields.