This paper proposes the design of optimal wind farm configuration using a new wind probability distribution map at Huasai district, the east coast of Southern Thailand. The new wind probability distribution map integrates both frequency of wind speed and direction data at a monitoring site. The linear wake effect model is used to determine the wind speed at downstream turbines for the total power extraction from a wind farm array. The component cost model and learning curve is used to express the initial investment cost, levelized cost and the annual energy production cost of a wind, farm, depending on the number of wind turbines, the installed size, hub height and wake loss within a wind farm. Based on Thailand wind energy selling price consisting of the fixed wind premium on top of base tariff, the profit depends on revenue of selling electricity and cost of energy. In this paper, Binary Particle Swarm Optimization with Time-Varying Acceleration Coefficients (BPSO-TVAC) is proposed to maximize profit subject to turbine position, turbine size, hub height, annual energy production, investment budget, land lease cost, operation and maintenance cost and levelized replacement cost constraints. Test results indicate that BPSO-TVAC optimally locate wind turbines directly facing the high frequent wind speed and direction, leading to a higher profit than the conventional wind farm layout. (C) 2015 Elsevier Ltd. All rights reserved.