Searching the global optimal value in heat exchanger network synthesis (HENS) becomes more difficult along with the increasing scale of cases.In structural opitimization model,such as stage-wise superstructure,solution domain can be enlarged by adding the stages of network,which would involve multiple integer and continuous variables in optimization and result in a decreasing computational efficiency.Thus,a new nodes-based model with stream splits is proposed,which can promote the quality of result and decrease time consumption at the same time.The characteristic of this model is that it changes the matching mode of heat exchangers by setting the number of nodes on streams to quantify the positions of heat exchangers.The proposed nodes-based model has more flexibility in generating and eliminating the heat exchangers and strong ductility in exploring the solution domain.The random walk with compulsive evolution is modified for enhancing its searching ability and applying modified random walk algorithm with compulsive evolution (RWCE) into the new proposed model.The obtained results of Case 1,Case 2 and Case 3 are superior to those reported in the literature,which decreases 106 USD·a-1,1253 USD·a-1,23 444 USD·a-1,respectively,verifying the robust effectiveness of this new model in solving HENS problem.