Vehicle Routing Problem is key step in logistics distribution; excellent distribution strategy helps enterprises in improving service experience and reducing service cost. In logistics distribution system for cigarette, huge demand point amount will always be faced with, and customers will put forward requirements for time quantum of accepting service; this Paper researches cigarette distribution routing optimization problem within this big range. By combining actual demands, establish mathematical model for cigarette logistics routing optimization, and use genetic algorithm to solve this problem; propose genetic algorithm-based hybrid genetic algorithm using hill climbing algorithm for local optimization; carry out Matlab simulating calculation through algorithm case; compared to common genetic algorithms, new hybrid algorithm can increase algorithm convergence speed and also improve globally optimal solution quality of algorithms.