We propose a vision based control system to control the tip position of a tri-layer conjugated ionic electroactive polymer actuator (EAP). The position information of the EAP actuator is extracted from the visuals obtained via a simple webcam by using threshold and closing algorithms. The location information is used as the feedback signal to the discrete-time PID controller which is supported by Smith Predictor. The Smith predictor method is used to overcome the effects of the delay caused by the image processing on the control performance. Accuracy of the measurements obtained via the camera is then compared to those obtained via a laser displacement sensor. Results show that vision based control provides accurate positioning of the polymer actuator which may widen applications of this class of actuators.