In the modern era of computer and technology, images and videos play a vital role. Therefore, there is always a need for robust skin detection system in order to cope with the intolerable and objectionable contents. In this paper, an efficient method has been implemented for skin detection, which detects the skin in different images under different environmental conditions. We have used the two machine learning approaches i.e. Random Forests and Multilayer perceptron for skin detection. We have also then combined the results of these two approaches used. We have used total of 554 images in our experiments.