The operator of a modern agricultural machine is unable to physically see all around the machine, a factor that contributes to accidental run-overs. There is a need to devise an effective blind-spot detection system for agricultural machines to enable operators to avoid accidental run-overs. The purpose of the study is to identify blind spots around two specific tractors and then to propose a conceptual blind-spot detection system based on the observed locations of blind spots. Grids were constructed around all four sides of the tractors to assess blind spots. The Human Activity Assistive Technology (HAAT) model was used to identify the requirements for a blind-spot detection system from the operator’s perspective. Diagrams were created to display blind spots around the two tractors. A camera-based system, which was able to eliminate all blind spots, was proposed as a feasible blind-spot detection system.