The problem of delimiting permanent basic farmland is essentially a multiobjective optimization problem. The traditional delimitation method can not take into account the quality and continuity of permanent basic farmland, and it can not balance the relationship between agriculture and urban development. A multi-objective permanent basic definition model was proposed based on particle swarm optimization algorithm. This model defined the general rules for delineating permanent basic farmland, the formal expression of delineation goals and constraints, and extracted three sub-goals—land suitability, land continuity, and land stability. At the same time, the model took into account four types of constraints, including the total area of permanent basic farmland, land use, terrain basic conditions, and town boundaries. It also described the coding and initialization methods of the algorithm, the particle position and speed update mechanism, the fitness function design, and the overall process. Finally, Xun County, Henan Province, was selected as the research area, and set up control experiments according to different target preferences. The results showed that the model was operable in the study of the delineation of permanent basic farmland, which can effectively reduce the subjectivity and arbitrariness in the delineation process, making the delineation of permanent basic farmland more scientific and reasonable. This would be conducive to improving the quality and production capacity of cultivated land, thereby ensuring national food security and ecological security.