In order to accurately and quantitatively obtain an industrial layout scheme of an irregular port-vicinity industrial zone, the harbor area and the target industrial land were divided into square grids of the same size by the grid method in this study. The relevancy degree between grids is determined by the known relevancy degree among industries, harbor areas and industries, and the irregular port-vicinity industry with the goal of maximizing the comprehensive relevancy degree among all grids isconstructed. Finally, an integer-coded multi-agent evolutionary algorithm is compiled based on grid number sorting to obtain the layout scheme with the greatest comprehensive correlation. Taking the Yancheng Binhai Port as an example, the coastal port industry is divided into 50 grids. According to the planned area, the grids occupied by six industries to be developed are determined. The multi-agent evolutionary algorithm is compiled by MATLAB. The agent grids are set to 20×20 to run 10 times for 150, 200 and 300 generations of evolutionary algebra, and three layout schemes with the greatest correlation degree are obtained. The comprehensive correlation degree is 549, 550.8 and 550 respectively. The best result of running 50 000 generations of genetic algorithm is only 520.8, which shows that the model and algorithm have strong superiority in solving the problem of irregular layout of port-vicinity industrial zones quantitatively.