To describe the interaction and evolution rule of network travel time and route choice probability, the paper proposes a quasi-dynamic model with the road section flow state and logit principle. The diverge-merge model and function between speed and density is used to compute section storage vehicles and travel time of non-incident section by analyzing traffic flow evolution process of incident section. The Kinematic wave theory is adopted to compute queue length and travel time of incident section. The results indicate that: before incident, traffic flow and route choice probability approaches an equilibrium with a long-time simulation; during the incident period, because of the interactions among queue length, route travel time and route choice probability, these three variables present concussed. With the incident clearance, route travel time decreases continuously; after queue dispersing, the route travel time and route choice probability approaches a new equilibrium within certain time.