Five phases algorithm (FPA) is one of the most effective algorithms for solving global optimization problems. The search operations consist of generating and overcoming strategy, and learning strategy. The algorithm has a simple structure and no predefined parameters, providing flexibility for practitioners. This paper suggests an enhanced form of FPA, named eFPA. The purpose is to prevent the algorithm from getting stuck in local optima and premature convergence in complex problems. We combine the fitness-distance balance method to redesign the generating and overcoming strategy. The designed algorithm is tested using the CEC 2022 and compared with the nine advanced algorithms. The statistical test results show that the proposed eFPA is highly competitive and explores the search space better than the compared algorithms.