In this paper, under the complex network framework, we study a seasonal influenza-like disease model by incorporating the interplay between subsidy policies and human behavioral responses. In the model a small proportion of individuals are freely vaccinated according to either the targeted or random subsidy policy in advance, while the remaining individuals choose to vaccinate (or not) based on voluntary principle and update their vaccination decision via an imitation rule. Our findings show that the targeted subsidy policy is only advantageous when individuals prefer to imitate the subsidized individuals’ strategy. Otherwise, the effectiveness of the targeted policy is worse than that of the random subsidy policy, since individuals preferentially select non-subsidized individuals as their potential imitation objects. More importantly, under the targeted subsidy policy, preferential imitation causes a non-trivial phenomenon: that the final epidemic size increases rather decreases with the proportion of subsidized individuals. We further define social cost as the sum of the costs of vaccination and infection, and study the impact of each subsidy policy on the social cost. Our result shows that there exist some optimal intermediate regions leading to the minimal social cost.