OBJECTIVE : To determine whether two-dimensional video analysis could discriminate running kinematics between recreational runners with and without a running-related knee injury. METHODS : Case-control. METHODS : Research laboratory. METHODS : Forty-two recreational runners (5 male-13 female injured; 7 male-17 female non-injured). Running-related knee injury was defined as the presence of anterior or lateral knee pain, resulting in altered running activity for at least one week. METHODS : Foot and tibia inclination at initial contact, and lateral trunk position, contralateral pelvic drop, femoral adduction, hip adduction, knee flexion and ankle dorsiflexion at midstance were measured with two-dimensional video analysis during running. Participant characteristics (sex, age, body weight, body length, body mass index, running volume before injury, running speed) and two-dimensional measured angles were compared between groups. RESULTS : No significant differences in participant characteristics between groups were identified (P > .05). The injured group ran with greater contralateral pelvic drop (P = .035), femoral adduction (P = .021) and hip adduction (P = .001) at midstance, and significantly smaller foot inclination at initial contact (P = .031). CONCLUSIONS : Two-dimensional video analysis can discriminate kinematics between runners with and without running-related knee injury. Greater contralateral pelvic drop, femoral adduction and hip adduction at midstance may provide running retraining targets for runners with running-related knee injury.