Abstract: In order to find the best combination of transport modes in an intermodal network ,a genetic algorithm is presented and the key connection joints in a city are coded into integer numbers instead of the traditional binary digits . This can reduce the length of the code and simplify the coding work .The collocation of transport modes at the connection joints and the transfer of transport modes can be studied by dynamic programming .The intermodal network data are stored by array of structure and linked storage structure based on the multi-graph structure .This method is found to be feasible in simulation ,capable of finding the optimum solutions in a very short time and with a small number of iterations .