To achieve a reasonable level of precision in tractor-based field operations, a tractor operator has to guide accurately, monitor and control both the tractor and the attached implement. Since guidance is the most time consuming task among the others, researchers have attempted to automate the guidance task. However, the use of automatic guidance and control in agricultural applications is not always appropriate. Transportation of the vehicle on a public road is an example of this. Some researchers therefore have focused on Vision-Aided methods to give some guidance aid to the driver rather than on eliminating the driver. To investigate the accuracy of such methods, a Vision-Aided tractor guidance belt-type simulator was developed. An experimental prototype of the simulator was constructed. To evaluate the prototype, a completely randomized factorial experiment was conducted with forward speed, heading angle, and camera tilt angle being the major factors under investigation. The simulator performed satisfactorily at 5 and 7km/h and mean deviations of 1.14 and 2.31cm were obtained respectively.