Abstract: Loading-transportation plays an important role in a distribution system and is crucial to its efficiency.With the complex constraints of distribution system considered,an integrated planning model for goods loadingtransportation with single source and multi path is developed.The problem is then solved by using an improved genetic algorithm with techniques such as serial number coding,maximum reserved crossover,and selfadaptability.The feasibility and effectiveness of the proposed method are demonstrated by an example.The problem of loadingtransportation with single source and multi path can be effectively solved.This is a new way for optimization of logistics distribution system.