The resilience of buildings plays a crucial role in reducing repair costs and minimizing operational disruptions after disasters. Among the various strategies developed to mitigate seismic effects, passive control systems are widely recognized, with Tuned Mass Dampers (TMDs). This research focuses exclusively on TMD systems and aims to determine their optimal parameters which are mass ratio, frequency ratio and damping ratio, to minimize superstructure displacement under seismic loading. To achieve this, the optimization process employs the Frequency Response Function (FRF) as the cost function and leverages Genetic Algorithms (GA) to handle the inherent non-convexity of the optimization problem. The GA methodology is chosen for its robustness and efficiency in finding globally optimal solutions, outperforming traditional analytical or heuristic approaches often found in the literature. The effectiveness of the optimized TMD system is evaluated through time-history analyses, using recorded earthquake data from the PEER database. A six-story reinforced concrete building is used as a case study, with its uncontrolled response serving as a baseline for comparison. The results clearly demonstrate that the GA-optimized TMD achieves significant reductions in structural acceleration response by 88, 77, 94, 87, 79 and 88%, underscoring its potential to enhance building resilience under seismic conditions. This study highlights the superiority of Genetic Algorithms in identifying the optimal properties of TMD systems, offering a reliable framework for improving seismic performance in practical applications.