Based on the blade layers, a multi-criteria constrained optimum design model for wind turbine blades is developed and programmed. This model is pursued with respect to minimum blade mass to reduce the cost of wind turbine production. Because the spar caps are the main parts to endure the loads in modern wind turbines, the thickness and the location of the layers on spar caps are chosen as the optimization variables. During the process of design, a lot of criteria need be satisfied. The maximum blade tip deflection is one of the most important criteria in the blade design. However, to get it, at least 130 load cases need be computed in engineering design, so the process is very time consuming. To decrease the computation time, the load case which presents the maximum tip deflection in the initial design is selected and calculated by FAST software in this model. Besides, an improved particle swarm optimization (1350) algorithm, which has better optimization capability and efficiency than the original PSO algorithm, is used to search the optimum solution. To build this model based on the blade layers, a key module used to link the FAST and the improved PSO algorithm is programmed and named PreLayers. At last, this optimization model is applied in an MW wind turbine design. The optimal results have also been tested by FOCUS5. The computed results show that this model is efficient and can be an optimal design tool in the engineering design. (C) 2011 Elsevier Ltd. All rights reserved.