There are so many vehicles in the world and the number of vehicles is increasing rapidly. To alleviate the parking problems caused by that, the smart parking system has been developed. The parking planning is one of the most important parts of it. An effective parking planning strategy makes the better use of parking resources possible. In this paper, we present a feasible method to do parking planning. We transform the parking planning problem into a kind of linear assignment problem. We take vehicles as jobs and parking spaces as agents. We take distances between vehicles and parking spaces as costs for agents doing jobs. Then we design an algorithm for this particular assignment problem and solve the parking planning problem. The method proposed can give timely and efficient guide information to vehicles for a real time smart parking system. Finally, we show the effectiveness of the method with experiments over some data, which can simulate the situation of doing parking planning in the real world.