It is found that most biases of tropical rainfall and clouds in long-term climate simulations are also shown in 5-day hindcasts initialized with reanalyzed observations (Xie et al. 2012, Ma et al. 2013, 2014). Here we develop a systematic, diagnostic method, by using a set of Transpose Atmospheric Intercomparison Project II (Transpose-AMIP II) simulation with CAM5 during the winter of 2009-2010, to analyze the possible physical mechanisms responsible for the precipitation biases over tropical oceans. Particularly, the diagnostics will be focused on the linkage between the representations of physical parameterization in the model and the fast, moist-related processes in the tropics, and also the large-scale moisture transport in the Tropics. We stress the importance of the representations of deep-convection and low-clouds in lower-tropospheric moisture budget which is related to the tropical rainfall. We also investigate the influence of both the cloud radiative effect (CRE) and the low-level-water-vapor-induced downward long-wave radiation on the tropical near-surface temperature biases, which in long-term may influence the modeled tropical rainfall.