Objective: It was aimed to reveal the foot shape variations in three dimension via the 3D statistical shape modelling. Methods: Fifty feet of normal Chinese young males were used build the statistical shape model. Steps, including mesh registration of foot surface, axis alignment and principal component analysis (dimension reduction), were performed to obtain the parameterized foot shape (mean shape and standard deviation of principal components). Results: Through the principal component analysis (PCA), the 3D foot shape varies in the length and width (PC1, 48.01%), arch and dorsal height (PC2, 11.38%), and hallux abduction-adduction position (PC3, 7.48%). Conclusions: Based on the parameterised 3D foot statistical shape modelling, these dataset could be applied into the population-based shoe last manufacture, orthotics customization and quick diagnosis of foot disorders in clinics.