Hair is a vital component of a person’s identity, and it can provide strong cues about age, background and even personality. More and more researchers focus on hair modeling in the field of computer graphics and virtual reality. Traditional hair modeling method is physically-based simulation by setting different parameters. The computation is expensive, and the construction process is not intuitive and difficult to control. The image-based hair modeling method has the advantages of fast modeling and high fidelity. In recent years, researchers have attached great importance to this method. This paper reviews the development of hair modeling based on single image, static hair modeling based on multiple images, dynamic hair modeling based on captured videos, and the editing and reusing of hair modeling results. It also analyzes and summarizes the application and inadequate of each method. In the first section, it summarizes the single-image based hair modeling method which can be divided into two types of orientation-field-based and data-driven-based hair modeling method. The static hair modeling and dynamic hair capture methods are reviewed in the next two parts. The static hair modeling based on multiple images also can be divided into two types of orientation-field-based and data-driven-based method. In the last section, the editing and reusing of hair modeling results are reviewed. The future development trends and challenges of image-based hair modeling are proposed in the end.