A double-moment bulk microphysics scheme, Milbrandt 2-mon (MY), was evaluated using the WRF single- column model during the period of the Tropical Warm Pool International Cloud Experiment (TWP-ICE). Results from separate simulations using the double-moment and sample one-moment (1M) versions of the microphysics scheme with the default settings in WRF were able to reproduce the general characteristics of moisture variables from the macro to the micro scale, as compared with observations and cloud resolving model (CRM) results. The rain rate and microscale features of the ice phase from the double-moment control simulation (CTL) were very close to observations, but showed significant deviations from the sample 1M version simulation. One-moment bulk microphysics schemes have been widely applied in real business situations with little computation, but their simulating abilities need to be improved. In order to maintain a reasonable computational expense, but with better simulating ability, several measures were applied to improve the sample 1M versions of the MY scheme on the basis of the microscale features of the ice phase from CTL. The results from the runs with two ice 1M modified versions were similar. Both showed non-significant improvements, despite their different treatments in terms of the number concentration of ice crystals. The frozen water path and observations showed more agreement, but significant changes occurred in the composition of ice clouds, mainly attributed to the weakened growth of snow and graupel indirectly affected by smaller ice crystals. The liquid water path showed an order of magnitude larger than observations as a result of abnormal enhancement in cloud droplet content, related to enhancements in upward movement. The liquid water content and raindrop effective radius changing in an opposite direction led to a tiny improvement in the rain rate. Sensitivity tests using two snow 1M modified versions of the MY scheme showed that ice-phase cloud was enhanced after adding the snow intercept and improving the environmental conditions of superstation. The characteristics of the snow intercept, liquid water content, and rain rate became significantly improved after adjusting the empirical formula of the snow intercept properly. Meanwhile, the method of adopting the reference values of the snow intercept in CTL did not show the same performance in ice-phase clouds, resulting in thicker warm clouds and a significant discrepancy between the simulated and observed rain rate. Optimized simulations with 1M modified versions suggested that the treatment of adjusting the empirical formula of the snow intercept properly in the 1M version of the microphysics scheme has a certain feasibility.