Abstract: Human mobility is a cross-disciplinary research hotspot which reflects the complex man-land relationship. Understandings of the intra-urban employment population flow, as an important part of urban group mobility, are crucial for urban planning and traffic forecasting. The common use of location-awareness devices such as mobile phones enables to capture human behavioral data for analyzing intra-urban employment flow. In this paper, we attempt to characterize the employment flow and interpret its formation mechanism using mobile phone data recorded during 30 days (from June 1, 2016 to June 30, 2016) in Wuhan. Considering the spatial distribution of cellular base stations, we divide our study area into grids sized 250 m×250 m. We then adopt approaches to extract daily individual mobility information by identifying positon and duration of stagnations. Next, we propose several rules to infer in which grids people work and reside from individual mobility information we extracted and estimate the job/housing distribution as well as the employment flows between sub-districts. We then build the employment flows network, and methods such as network centrality analysis, network density analysis and community detection are applied to discover the pattern and characteristic of employment flows. Finally, we examine the influence of employment potential, traffic accessibility, spatial proximity, cultural difference and major industry on the size of employment flow between sub-districts and interpret its formation mechanism using logistic regression. The results firstly show that the number of intra-urban employment flow in Wuhan is unevenly distributed as large amounts of employment flow are concentrated in a few sub-districts. Secondly, the employment flows mainly occur in the sub-districts with good spatial proximity. It shows the distribution of the outflow employments decrease gradually as distance and accessibility time increase, and forms several employment communities (such as Hanyang, Wuchang North, Wuchang South, et al.) in the influence of geographical barriers and cultural differences. Thirdly, the employment potential measured by the resident population of the origin sub-districts and the working population of the destination sub-districts is the most important factor to increase the size of employment flow. Rather, cultural differences and poor accessibility hinder employment flow. The result also verifies that differences in major industries affect the employment flow, that commerce and education have a negative effect on the outflow of employments, but the industry attracts employments.