Population connectivity among 7 fishery grounds of Ezo abalone Haliotis discus hannai on the northern Miyagi coast (approximately 80 km) was investigated using hydrodynamic and particle-tracking models. The objectives were to: (1) clarify the larval dispersal processes, (2) quantify the dispersal distance of larvae and (3) estimate the population connectivity. To simulate larval dispersal, particles were released at the timings of spawning estimated from the shell lengths of newly settled abalone. The larval dispersal was simulated for 2 periods. The modeled hydrodynamics in the first period simulated stormy conditions because of the passage of a low-pressure system whereas that in the second period simulated relatively calm conditions. In the first period, the spawning appeared to be triggered by the low-pressure system, and the larval dispersal was estimated to be generally greater than that in the second period. The mean dispersal distances were less than 40 km in both periods. Model results indicate that abalone in the fishery grounds on the mid- and southern coast exhibit 2 distinct larval dispersal patterns. The number of connected sites was 3 to 7 and 2 to 4 in the stormy and relatively calm conditions, respectively. The calmer hydrodynamic conditions were favorable for self-replenishment. Self-recruitment was usually greater than the connectivity with the other fishery grounds; however, poor self-recruitment occurred in some fishery grounds. This implies that the restocking and protection of local resources do not always lead to an increase in local recruitment, and it is necessary to investigate the larval dispersal processes from each larval source for successful management.