Monitoring the main compositions of milk content like fat, lactose, protein and total solids, has become a major challenge in dairy cattle farming. For quantitative determination of fat content in milk based on the relation of milk color features different methods have been used, but long time, high cost, and need for experts for analysis are some disadvantages of them. In this study, for rapid monitoring of milk fat content, novel technology of image processing coupled with artificial neural network (ANN) and Particle swarm optimization (PSO) methods has been applied. The estimated milk fat content of the best proposed method was extensively compared with the reference sample (R2=0.99, MAE=0.22, and MSE=0.05). Moreover, effect of water on color components of milk with different percentages of fat content have been investigated. Results approved the proposed method as a reliable, rapid and low-cost method for monitoring milk fat content.