Skin-color modeling is a crucial task for several applications of computer vision. Problems such as face detection in video are more likely to be solved if an efficient skin-color model is constructed. Most potential applications of skin-color model require robustness to significant variations in races, differing lighting conditions, textures and other factors. Given the fact that a skin surface reflects the light in a different way as compared to other surfaces. As the color of human skin is created by the combination of blood (red) and melanin (brown, yellow) which gives it a restricted range of hues. A skin region can be classified by comparing large image content of skin database and non-skin database. The RGB color space is widely used and most effective to detect skin region from an image. The segmentation is used to localize and identify homogeneous regions in a picture by perceptual attributes which include the size, the shape and the texture and/or color information. The probability of each RGB color space of skin and non-skin database is important to detect skin pixels. For each pixel of testing image, the RGB value is calculated then the probability ratio of that RGB color space for training Skin and non-skin data base is compare with a threshold variable called . The threshold value lies between 0 and 1 but for this analyzing purpose threshold value has been taken as 0.4. If the ratio will be greater than 0.4 then that pixel will be detected as skin pixel elsedetected as non-skin pixel. After segmenting out skin from images, this can be useful for identifying faces, hand sign recognition, offensive content such pornography. The performance curve (ROC curve) reflects the overall accuracy of our analysis.