As the most basic travel mode of residents, walking trip takes place more frequently in residential districts. It is of great significance to study the land use characteristics of the residential district suitable for walking in order to create a good walking environment and land use model. In this paper, 18 residential blocks in Xi'an are taken as the research object, based on the relevant data of the pedestrian flow rate and land use of residential blocks, this paper analyzes the interactive relationship between the non-commuting hours during the week, the commuting hours during the week and the weekends by using the multiple linear regression model, and the land use characteristics of residential blocks under the guidance of walking trip are revealed. The results show that the land use factors related to pedestrian flow rate are mainly 9 factors, such as building density, buildings mixing degree, mixing degree of business functions and density of functional formats. In different hours, walking trips have different demands on various elements of land use. During the non-commuting hours during the week, the demand for walking trips for building density, mixing degree of business functions, density of functional formats and footpath area ratio is large. During the commuting hours during the week, the demand for walking trips for buildings mixing degree, density of functional formats, road network connectivity, and green area ratio of footpath is large. During the weekends, the demand for building density, density of functional formats, road network density, and footpath density is large. Through the above analysis, the optimization strategy is proposed for residential blocks in Xi 'an, which provides a theoretical basis for the optimization and transformation of the space environment of residential blocks to promote walking trip.