The modern human foot is a complex structure thought to play an important role in our ability to walk and run efficiently. Comparisons of our feet to those of our evolutionary ancestors and closest living relatives have linked the shape of several foot components (e.g., the longitudinal and transverse arches, size of the heel and length of the toes) to specific mechanical functions. But since foot shape varies widely across the modern human population, this study aimed to investigate how closely foot shape, deformation and joint mechanics during various locomotor tasks are actually linked. And whether the latter can be accurately predicted based entirely on the former two. A statistical shape-function model (SFM) was constructed by performing a principal component analysis on 100 participants' three-dimensional foot scans, and joint angles and moments captured during level, uphill, and downhill walking and running. This SFM revealed that the main sources of variation were the longitudinal and transverse arches, relative foot proportions and toe shape along with their associated joint mechanics. However, each of these only accounted for a small proportion of the overall variation in foot shape, deformation and joint mechanics, most likely due to the high structural complexity and variability of the foot. Nevertheless, a leave-one-out analysis showed that the SFM can be used to accurately predict the joint angles and moments of a new foot based only on its shape. These results have implications and potential applicability across numerous fields, such as evolutionary anthropology, podiatry, orthopaedics and footwear design.