Partial face images are produced in an unconstrained environment. A face may be occluded by sunglasses, a hat and a scarf, captured in various poses, positioned partially out of cameras field of view. Human face plays an important role in our social interaction, conveying people’s identity but it is a dynamic object and has a high degree of variability in its appearances. The problem of recognizing an arbitrary patch of a face image remains largely unsolved. This study proposes a new partial face recognition approach, called Dynamic Feature Matching, which combines Fully Convolutional Networks and Sparse Representation Classification to address partial face recognition problem regardless of various face sizes. DFM does not require prior position information of partial faces against a holistic face.