There are some problems in group distributed manufacturing enterprises, such as the scatter of geographically location, the mismatching of manufacturing resources and abilities, the coexistence of resources idle and shortage etc. Based on the characteristics of multi-agent, multi-task, multi-resource, multi-process and co-ordination in the process of manufacturing resource allocation, the optimal model of distributed manufacturing resource allocation is proposed to balance the overall interests of the group and the individual interests of the sub-ordinate enterprises. An improved genetic algorithm based on Logistic chaos is designed to obtain the Pareto optimal solution of the model. Finally, an example with its analysis is given to demonstrate the effectiveness of the proposed model and algorithm.