The purpose of this study was to explore the relationship between variability in pitching kinematics and consistency in pitch location. Data were collected for 47 healthy baseball pitchers throwing ten full-effort fastballs to the centre of the strike zone. For each pitch, 20 kinematic parameters were calculated with an automated motion capture system while pitch location was measured with a PITCHf/x system. Variability of each kinematic parameter was defined for each pitcher as the standard deviation among his fastballs thrown. For calculating consistency, each pitcher’s mean pitch location was first calculated. The distances from each individual pitch to the mean pitch location were then found for each pitcher tested. A consistency metric was then calculated for each pitcher by averaging these distances. A multiple linear regression model was developed using stepwise regression with backwards elimination. The resulting model explained 58% of the variance in the consistency metric and included five parameters, three at foot contact (upper trunk tilt, shoulder abduction, and shoulder horizontal abduction) and two at time of maximum shoulder external rotation (shoulder external rotation and shoulder horizontal adduction). Reducing variability at the shoulder during the early portions of the pitching motion may improve consistency of ball location.