The purpose of this study is to identify homogenous subgroups of foot-ankle (FA) kinematic patterns among recreational runners and further investigate whether differences in baseline movement patterns can influence the mechanical responses to a foot-core exercise intervention program. This is a secondary analysis of data from 85 participants of a randomized controlled trial (clinicaltrials.gov – NCT02306148) investigating the effects of an exercise-based therapeutic approach focused on FA complex. A validated skin marker-based multi-segment foot model was used to acquire kinematic data during the stance phase of treadmill running. Kinematic features were extracted from the time-series data using a principal component analysis, and the reduced data served as input for a hierarchical cluster analysis to identify subgroups of FA movement patterns. FA angle time series were compared between identified clusters and the mechanical effects of the foot-core exercise intervention was assessed for each subgroup. Two clusters of FA running patterns were identified, with cluster 1 (n = 36) presenting a pattern of forefoot abduction, while cluster 2 (n = 49) displayed deviations in the proximal segments, with a rearfoot adduction and midfoot abduction throughout the stance phase of running. Data from 29 runners who completed the intervention protocol were analyzed after 8-weeks of foot-core exercises, resulting in changes mainly in cluster 1 (n = 16) in the transverse plane, in which we observed a reduction in the forefoot abduction, an increase in the rearfoot adduction and an approximation of their pattern to the runners in cluster 2 (n = 13). The findings of this study may help guide individual-centered treatment strategies, taking into account their initial mechanical patterns.