In the past, many approaches have been devised to address the load balancing problem for social networked multiagent systems (SN-MASs). However, few of these approaches consider the migration cost incurred when migrating tasks for load balancing, moreover, current SN-MASs often consist of communities, and the migration costs of intra-community and intercommunity transfers are heterogeneous. To minimize the load imbalance of agents and to incur the least migration cost, this paper introduces a net profit-based load balancing mechanism. In this mechanism, each load balance process (i.e., migrating a task from one agent to another agent) is associated with a net profit value which depends on the benefit it gains by making a contribution to alleviating the system load unfairness and the cost of migrating the task. The agents always perform the optimal load balance process that has the maximum net profit value, thereby improving system performance, as well as reducing the migration cost. Our simulations show that our approach not only guarantees that agents can undertake fair loads but also reduces the overhead migration costs compared with the previous load balancing approaches that ignore the cost of migrating the task.