Detecting and interpreting seismic waveforms is of utmost importance in seismology, providing means for the early hazard estimation that can give rise to both the increase of rapid earthquake detection systems and strategies for countermeasures. Traditional methods frequently require manual analysis, which is less efficient and potentially more imprecise. Genetic Algorithms (GAs) have been proven to be a versatile optimization technique that can be used in seismic data analysis and classification tasks. This study focuses on optimization of models such as Logistic Regression and Decision Tree by employing Genetic Algorithms for classification analysis of seismic waveforms, based on Stanford Earthquake Dataset (STEAD). This study aims to improve these models using GAs. Several studies illustrate the capability and performance of GAs for different seismic or classification tasks, possibly indicating a positive transition across domains to improve learning. This study fills a void in the existing literature by specializing the use of GAs for seismic waveform classification and qualitatively evaluating GA-optimized models’ effectiveness. The results are anticipated to help improve the accuracy and speed of earthquake monitoring, advancing seismic data analysis.