With the rapid development of mobile internet technology, new requirements have been set for the mobile terminal positioning method. In view of the problem that we cannot take into account of energy consumption, accuracy and versatility, we propose a Hidden Markov Model-based and personal habits-based network location algorithm. The algorithm divides the roads into sections in the area covered by the base stations according to the map, and records the personal walking trajectory, which helps build the hidden markov model of GSM signals with the statistical law. Thus, with only a station’s signal sequence, we can get an accurate estimate of the user’s current position. Experiments show that the algorithm can guarantee a good positioning accuracy with lower energy consumption.