Subdomain technique is a computationally efficient approach with an acceptable accuracy for performance evaluation of the electric machines. This paper develops the subdomain model of three types of permanent magnet (PM) machines including surface-mounted PM (SPM), Inset PM, and consequent pole PM (CPPM) machines. The performance parameters of the studied machines are extracted via subdomain technique and validated by 2D finite-element analysis (FEA). The airgap flux density is calculated using the vector potentials obtained from the subdomain model. The flux density of the iron parts, that is required to find the core losses, are found using the magnetic vector potentials in the iron parts. The developed model is then employed for optimal design of various wind generators. The optimization is executed through two different scenarios. Firstly, the machines are optimized for maximization of the generator efficiency for a single operating point. In the second scenario, the optimization is done over annual wind regime of an area to maximize the annual energy efficiency. The airgap length, magnet dimensions, stack length, and the rotor and stator diameters are the variables of the optimization problem. The optimal designs are compared to show the advantageous and disadvantageous of the considered machines for the wind generator application.