Abstract: Based on the anonymities provided by the service providers, accumulated mobile positioning bit data, the travel information of nationwide population during the spring festival is extracted, population mobility situation in the epidemic outbreak area is based on to construct model to calculate the number of infected individuals in Wuhan and the increasing trends. The study results show that before the lockdown of Wuhan, based on the infected status of 50 cities outside Hubei province, DPNA-M model is built to estimate the number of infected people in Wuhan to be 26 464, the result is close to early overseas estimate results. Compared with existing estimation methods, the model is not strictly dependent on the statistics data in the epidemic area, its estimation results are more accurate, more superior in the stability and application universality of model.