Hot Isostatic Pressing Silicon Nitride (HIPSN) full-ceramic ball bearings have excellent properties including light weight, wear resistance, and good accuracy retention, which are promising for application in high-end precision machine tool spindle systems. The service performance of spindle bearings directly determines the operational performance of precision machine tools. In order to improve its performance, it is necessary to optimize the design of full-ceramic angular contact ball bearings. In this study, the internal macrostructural parameters of the bearings are considered as design parameters, with the raceway wear rate and stiffness serving as the objectives. Based on the Advanced Dynamics of Rolling Elements (ADORE) software, response surface models are established using Response Surface Method (RSM) to characterize the correlation between design parameters and objectives. Optimization is performed using multi-objective genetic algorithm (MOGA) to obtain multiple sets of optimized solutions. Compared to the original bearing design parameters, the maximum reduction in raceway wear rate of 9% and the maximum increase in stiffness of 47.5% have been achieved in multiple sets of optimization results. The optimization results offer a theoretical basis for the implementation of high-speed precision full-ceramic ball bearings.