Abstract: For implementing wavelet transform in analog hardware systems with very low power consumption and small size, a particle swarm optimization method is employed in the approximating of wavelet functions. First, utilizing the least-squares error criterion, a general mathematical model for approximating wavelet functions is established. Then the technique of L2 approximation based on Particle Swarm Optimization (PSO) is presented, which is more attractive. These techniques are compared by means of a worked example, involving some wavelet approximation. The L2 approximation approach based on PSO is shown to exhibit superior performance.