Highway maintenance operations occupy existing road infrastructure and have a significant impact on vehicle traffic safety and efficiency. In order to analyze the traffic flow characteristics during maintenance operations on a four-lane highway when the inner lane is closed, a traffic flow modeling and simulation method based on an improved cellular automaton was proposed. According to the Highway Maintenance Safety Operation Regulations (JTG H30—2015), maintenance operation control zones were established, dividing the mainline highway facilities into six traffic scenarios: warning zone, upstream transition zone, buffer zone, work zone, downstream transition zone, and termination zone. By introducing a longitudinal safety distance model and optimizing the lateral lane-changing safety condition determination rule, the NaSch car-following model and symmetric two-lane cellular automata (STCA) lane-changing model were improved. The model was calibrated using field data, and the traffic flow operating conditions in the maintenance work zone were simulated using MATLAB. The results show that when the traffic flow density reaches 1 550 pcu/h per lane, setting the merging start point 1 000 meters downstream of the warning zone, with a speed limit of 60 km/h and an upstream transition zone length of 160 meters, the traffic flow safety and efficiency indices in the maintenance zone are optimized.