Face shape and hairstyle are two interrelated elements that will affect a person's overall appearance. The shape of the face has different shapes, such as round, square, oval, diamond, and heart-shaped, choosing a hairstyle must adjust the shape of the face in order to get the appropriate results. The problem that often arises when getting a haircut is choosing a hairstyle that does not match the shape of the face. The purpose of this research is to develop a system that can assist users in knowing the shape of the face and can provide recommendations for haircuts that match the shape of the face. This type of research uses the viola-jones method with the stages of haar-like features, integral images, AdaBoost Machine Learning, cascade classifier. The results of this study show that the system can detect the shape of the user's face and produce recommendations for haircut models based on face shape. The results of direct trials of the haircut model selection system based on face shape on 30 respondents resulted in 20% not suitable and 80% of the 30 respondents felt suitable with the results of selecting a haircut model based on face shape using the Viola-Jones method. The Viola-Jones method tested with the confision matrix obtained an accuracy value of 43%.