Abstract: A mathematical model of improved network maximum flow is established to study campus road planning in the research background of a domestic university. Firstly, according to the basic network maximum flow model, Ford-Fulkerson algorithm is used to solve the theoretical maximum capacity of 46 person/s. Secondly, considering the road selectivity, the shortest path model is established. Dijkstra algorithm is used to calculate the shortest path from each single source to each single sink, and A* algorithm is used to exclude the path with a large difference from the shortest distance, so as to screen out the effective path. Using the results of the shortest path model to strengthen the constraints in the original model, the simple method is used to solve the actual maximum capacity of 23 person/s. Finally, a linear programming model with the lowest cost of road widening as the objective function is established to optimize the road. The results show that the existing road design can meet the road traffic demand. If the road capacity needs to be improved and the change of the road is minimum, the key roads in the road network can be expanded appropriately.