This paper provides a comparative analysis of fault diagnosis in power systems using evolutionary computation and machine learning techniques. Power system faults, including line-to-ground, line-to-line, and three-phase faults, pose significant challenges to the stability and reliability of electrical grids. Efficient and accurate fault diagnosis is crucial for minimizing downtime and maintaining optimal power delivery. This research evaluates the performance of various evolutionary computation algorithms such as genetic algorithms, particle swarm optimization and differential evolution alongside machine learning methods including decision trees, support vector machines, and k-nearest neighbor, in diagnosing power system faults. The study compares the accuracy, speed, and reliability of these methods in detecting and classifying faults under different system conditions. The findings highlight the strengths and limitations of EC and ML-based fault diagnosis techniques in power systems, providing valuable insights for researchers and practitioners in the field. Key Words: Fault diagnosis, Power system, Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Differential Evolution(DE) , Support Vector Machine (SVM) , Decision Tree (DT), K-Nearest Neighbors (KNN).