The series robots have nonlinear dynamics with together uncertainties in joints’ forces and robot’s model. For the fuzzy control, two parameters of angular position and angular velocity are usually used as the linguistic variables. This method is similar the PD controller, and, consequently it cannot track a desired trajectory and it have a steady-state error. In the present paper the summation of errors is additionally used as a linguistic variable. Using three linguistic variables of error, error derivative and error integration for the angular position, trajectory tracking of a RR series robot is performed for its uncertain dynamic model. This method is similar the PID controller, and, consequently it can track a desired trajectory. Additionally, the presented fuzzy PID controller is not sensitive to the noise and uncertainty, and, it presents better results than a classic PID controller