Nowadays, the change of marketplace require manufacturing system to diversify from mass customization style to step-mass customization style, thus lead to the manufacturing value orientation change from product oriented to custom oriented, main manufacturing strategy change from cost, quality etc. to deadline, manufacturing principle change from individual to cooperation, manufacturing direction change from technology-oriented to innovation and human centered. The proposed holonic control system for operating an automated material handling system attempts to combine both the advantages of hierarchical (global optimization) and heterarchical (real-time response when changing conditions occur) control architectures. Due to its architecture, the holonic system can work both as a hierarchical system, during normal operation when no disturbances are present, and as a heterarchical system, when facing sudden disturbances. In this paper, the negotiation Holon which is map to agent strategy is applied to cooperative activity among holons in a Holonic society. A novel algorithm based on PSO and GA for the order task Holon allocation problem is presented. Simulation results show that the architect and negotiation model is effective to the task allocation problem, the algorithm have good performance on settling task optimization problem.