The three-phase voltage and current waveforms from a Power System (PS) are not considered pure sinusoids due to the presence of, among others, the harmonic distortion. This work presents an approach based on the Particle Swarm Optimization (PSO) method for the harmonic component estimation in a PS. PSO is a technique of search/optimization that models the social behavior observed in many species of birds, schooling fish and even human social behavior. The technique uses a population of particles to search inside a multidimensional search space. The objective of the PSO is to adjust the speed and position of each particle, seeking for the best solution within the search space. The results demonstrate that the method can precisely identify the harmonic components in the distorted waveforms and it shows considerable advantages if compared to the most common algorithm for this purpose, the Discrete Fourier Transform.