In this paper, a new automated machine supervised learning method for bar detection scheme in spiral galaxies based on the Nonnegative Matrix Factorization algorithm have been presented. Nonnegative matrix factorization has been introduced in this paper to detection bar in spiral galaxies, which is very easy to use, and gives us a good accuracy. Detection bar in spiral galaxies is the main objective of this research. WE describe an entirely automated method that extract feature from spiral galaxies and then automatically bar detection. The algorithm is trained using manually bared and non-bared images of spiral galaxies. The algorithm show that the bar in spiral images from the EFIGI catalog can be detected automatically with an accuracy of 97.3% with an average processing time of 0.37 s per galaxy compared to bar detection carried out by other authors and manually detection.