The objective of this study was to develop a method of muscle structure measurement based on the automatic analysis of muscle fibers, proximal fascias, and distal aponeurosis movements as revealed by a time-series of ultrasound images. This method was designed to detect changes in the length of muscle fiber movements, and its validity was demonstrated in a time-series of muscle movement, slow ankle dorsiflexion (10 degrees/s), by comparison to manual measurement. The results showed that, when this method was used, the changes in the length of the muscle fiber under slow muscle movement were smaller than those in manual operations by novice individuals. However, with the proposed method, it was possible to obtain a sufficient degree of validity and reliability for the changes in the length of the muscle fiber length compared with those in manual operations, since the correlation coefficients exceeded 0.8 which was tested by the linear regression. The proposed method suggests that automation reduces the errors caused by manual operations and makes the processing of data possible in an acceptable amount of time.