This paper used Gabor filter feature for face detection. we design filters for elicitation(extracting) facial feature from the local image using individualistic property of spatial locality and orientation individualistic of the Gabor filter. The vector based property of Gabor filter is used as input of classifier, which is FFNN(feed forward neural network) on a reduced feature subspace learned by an approach simpler than PCA. Large amount of calculation is required due to vector input having large values. so, we delete some rows and columns in order to reduce size of matrix to 1/3 of actual size. In spite of using PCA method save more time but it’s not the best way. Face detection & recognition technology is used in various fields like security system, video conferencing and identification .face distingustion is recently implemented in software. The aim of this paper is to develop a classifier based multi-layer perceptron neural network for face detection This face detection system starts with image preprocessing and then the output image is trained using Back propagation algorithm.