The aims of this study were to determine if isolated maximum joint torques and joint torques during a maximum polyarticular task (i.e. cycling at maximum power) are correlated despite joint angle and velocity discrepancies, and to assess if an isolated joint-specific torque production capability at slow angular velocity is related to cycling power. Nine cyclists completed two different evaluations of their lower limb maximum joint torques. Maximum Isolated Torques were assessed on isolated joint movements using an isokinetic ergometer and Maximum Pedalling Torques were calculated at the ankle, knee and hip for flexion and extension by inverse dynamics during cycling at maximum power. A correlation analysis was made between Maximum Isolated Torques and respective Maximum Pedalling Torques [3 joints x (flexion + extension)], showing no significant relationship. Only one significant relationship was found between cycling maximum power and knee extension Maximum Isolated Torque (r=0.68, p<0.05). Lack of correlations between isolated joint torques measured at slow angular velocity and the same joint torques involved in a polyarticular task shows that transfers between both are not direct due to differences in joint angular velocities and in mono-articular versus poly articular joint torque production capabilities. However, this study confirms that maximum power in cycling is correlated with slow angular velocity mono-articular maximum knee extension torque.