With the increasing demand of media-rich entertainment and location-aware services from people on the road, how to disseminate the multimedia data in large-scale Vehicular Ad-Hoc Networks (VANETs) efficiently and reliably is a pressing issue. Due to the high mobility, large scale, and limited contact time between vehicles, it is quite challenging to support the multimedia data dissemination in VANETs. In this paper, we first utilize a hybrid framework to model the VANETs to address the mobility and scalability issues. Then, we formulate a utility-based maximization problem to find the best delivery strategy and select an optimal path for the multimedia data dissemination, where the utility function has taken the delivery delay, Quality of Services (QoS), and storage cost into consideration. With rigorous analysis, we obtain the closed-form of the expected utility of a path, and then obtain the optimal solution of the problem with the convex optimization theory. Finally, we conduct trace-driven simulations to evaluate the performance of the proposed algorithm with real traces collected by taxis in Shanghai. The simulation results demonstrate the rigorousness of our theoretical analysis, and the effectiveness of the proposed solution.