To fully utilize the excellent structural performance in building construction and improve the seismic performance of the building, based on the transformation principle of subset optimization algorithm for optimization problems, a constrained subset simulation optimization algorithm suitable for optimizing the maximum displacement angle of multi ribbed composite wall panels is designed. The Bayesian algorithm is used to construct a restoring force model for multi ribbed composite wall panels. The constrained subset simulation optimization algorithm and resilience model are used to optimize the seismic performance of 4-layer multi ribbed composite wall panels. The results show that the section height and the equivalent slant support width of the continuous column for the 4-story multi ribbed composite wall panel change from discrete distribution to aggregation with the increase of iteration. Finally, the sampling is stable in the 9th floor. At this time, the section height of the continuous column is 230 mm, and the equivalent slant support width is 525. After optimization, the failure probability of both extreme displacement angle states has decreased. When the peak ground acceleration is 1.0 g, the optimized second limit state failure probability is less than 100%. When the peak ground acceleration value is between 0.2 g and 0.6 g, both limit states show a rapid upward trend. The constrained subset simulation optimization algorithm and Bayesian quantitative resilience model proposed in the research can effectively optimize the seismic performance of multi ribbed composite walls.