This paper improves the global exploration capability of the Competitive Swarm Optimizer (CSO) by introducing an adjacent distance matrix selection strategy and proposes an Adjacent distance Matrix-based CSO (AMCSO). The traditional CSO updates offspring individuals by randomly selecting two particle individuals for competition, which ignores the differences and similarities of particle individuals. To fully utilize the knowledge of and fitness landscape and particle swarm to guide the selection mechanism in competition, our proposed AMCSO adopts an adjacent distance matrix based on the Euclidean distance between every pairwise particle individual, which allows the selection of particle individuals with closer distance for competition and further ensures population diversity. Comprehensive numerical experiments were conducted on 10-D and 20-D CEC2022 benchmark functions to investigate the performance of AMCSO against the original CSO and four well-known metaheuristic algorithms (MAs), and experimental results and statistical analysis confirm the effectiveness of the integration of the adjacent distance matrix and demonstrate a statistically significant improvement of AMCSO.