This paper presents a new method using PSO for face detection, which does not utilize training samples. The method is based on the edge density of the image. In the preprocessing stage a face is approximated to a rectangle. Then PSO algorithm is applied to search for the best rectangle region. The rectangle area with best fitness value will be detected as the face region. Simulation results show that this PSO-based method is convergent and effective, especially for the case of images with the non complex background.