Abstract: Travel survey is an important data source for travel demand forecasting models.It is primarily relying on artificial completed,which is time-consuming and laborious.Besides,its data amount is too small to be used in actual situations.To solve these problems,a mobile phone APP for travel survey is developed and its application on traffic planning is developed.Currently,a person who stays indoor cannot be accurately located by mobile phone data.In this study,velocity and accuracy of trajectory data of mobile phone are used to identify whether the person stays or not.Then "stay points" are merged according to distance.People can use this APP to add their travel information on these "stay points".Compared with a travel survey in Qingpu district in Shanghai,the recognition accuracy of the APP is 96%.This method can objectively collect spatial location and time information of travel activities of people,and is of great significance to improve the accuracy of travel forecasting models.