With the development of GPS, Internet, and mobile devices, the map searching services become an essential application in people’s lives. However, existing map searching services only support simple keywords based location search, which neglect users’ complex search requirements, such as searching a hotel surrounded by station, shopping mall, cinema, etc. In this paper, we propose a map searching framework which maps users’ complex search requirements into a graph pattern match problem. In this framework, the user’s query requirements are mapped into a undirected graph, where the vertexes indicate the searched locations, and the edges represent the distance between connected locations. In this way, we propose a common pattern searching algorithm to give the top k matches groups and arrange them using a similarity. The similarity is defined by multi-model information, i.e., the graph structure, location categories, review stars, and review counts. Moreover, this method also allows user to input ambiguous and uncertain query condition with sketching query map. To adapt the method to large-scale data, we also filter the candidate groups by effective pruning methods. The evaluations on Yelp dataset demonstrate the proposed method is effective and flexibility in both certain and uncertain query graphs.