This study addresses the optimization of temporary shelter locations for earthquake preparedness in Lima, Peru. A multi-objective approach is implemented using two evolutionary algorithms, NSGA-II and SPEA-II, to identify shelter sites that maximize population coverage, prioritize areas with high seismic vulnerability and risk, and ensure proximity to essential services such as healthcare facilities and water sources. The analysis integrates geospatial data on 5,855 potential shelters, evaluating scenarios under tuned algorithm configurations. Comparative results indicate that while SPEA-II achieves higher and more stable average fitness values, NSGA-II consistently reaches higher maximum objective values and exhibits greater diversity across the Pareto front, leading to a more balanced spatial distribution of shelters. For this reason, NSGA-II was selected for district-level and shelter characteristic analyses. The overlap with official municipal designations is limited ( 6 %), revealing differences in prioritization and highlighting underserved high-risk districts. An interactive web-based visualization was developed to support decision-making, enabling exploration of Pareto-optimal solutions and their spatial implications. The proposed framework offers a replicable methodology for seismic hazard–exposed urban areas, integrating optimization, spatial analysis, and decision support tools.