In nature and human society, individuals’ spatial movement and game interaction are often governed by distinct mechanisms. However, most previous studies on evolutionary game dynamics with migration have treated the interaction and migration radii as identical, which ignores the range asymmetry. In this paper, we decouple these two radii by treating the migration perceptual radius r_\rmm and game interaction radius r_\rmi as two fully independent parameters, construct a spatial prisoner’s dilemma model with continuous migration in two-dimensional continuous space, and investigate how variations in these two factors affect the evolution of cooperation. We find that r_m has a pronounced non-monotonic effect on cooperation. When r_\rmm is excessively small, individuals lack directional guidance for migration and the cooperation level remains close to that of random mixing. At intermediate r_\rmm, defectors can efficiently track and infiltrate cooperative clusters, leading to a pronounced cooperation valley that is robust across defection temptations and interaction radii. In contrast, when r_\rmm increases to a moderate level, cooperators achieve directed aggregation and form stable structures, substantially enhancing cooperation. A further increase in r_\rmm enables defectors to follow cooperators’ movement signals at a large scale, breaking cooperative clusters and causing the cooperation level to decline again. Spatial snapshots and polar migration-direction diagrams reveal the microscopic mechanisms underlying these regimes. Furthermore, the game interaction radius r_\rmi plays a key regulatory role in cooperation evolution. An intermediate r_\rmi most effectively supports the formation of cooperative clusters, while an excessively large r_\rmi weakens network reciprocity. In the r_i-r_m parameter space, the cooperation level exhibits a non-uniform distribution with sharp boundaries between high and low cooperation regions, confirming the joint regulatory roles of both radii. In addition, the migration step size and population density interact with r_\rmm to jointly determine cooperation outcomes. The promoting effect of an optimal r_\rmm is prominent at low population density, whereas high density suppresses the cooperation level. These results provide a theoretical perspective on how movement behavior facilitates cooperation in real-world complex systems.