Real-time crash prediction models are being developed for motorways using traffic data to predict safety conditions. A rule-based approach was used to create a model that accurately predicts both crash and no-crash instances. Factors like traffic volume, average speed, and speed differences between loop detector stations contribute significantly to crash predictions. The model can predict 70% of crash occasions and 90% of no-crash instances accurately. These findings can help predict crash likelihood on motorways for better safety management.