This study clarified the association between acceleration and the rates of changes in spatiotemporal variables on a step-to-step basis during the entire acceleration phase of maximal sprinting. 21 male sprinters performed a 60-m sprint, during which step-to-step acceleration and rates of changes in step length (RSL) and step frequency (RSF) were calculated. The coefficients of correlation between acceleration and other variables were tested at each step. There were positive correlations between acceleration and the RSF up to the second step. Acceleration was positively correlated with the RSL from the 5(th) to the 19(th) step. At the third and from the 16(th) to the 22(nd) step and from the 20(th) to the 21(st) step, there was no significant correlation, but weak relationships were found between acceleration and the RSF and RSL. The results suggest that the acceleration phase can be divided into 3 sections, and for sprinting to be effective, it is important to accelerate by increasing the step frequency to the third step, increasing the step length from the 5(th) to the 15(th) step, and increasing the step length or frequency (no systematic relative importance of step length or frequency) from the 16(th) step in the entire acceleration phase.