Plastic gears are widely used by engineers in various engineering applications due to their reduced noise while working, low cost, lighter weight, chemical resistance, flexibility, and the ability to operate without lubrication. A disadvantage found in these drives is the thermal stresses induced during continuous operation. Incorporating cooling holes into the design of plastic spur gears can reduce the thermal stresses on the gears. These cooling holes promote increased stress and tooth deflection, thus exerting a negative effect. Various machine learning algorithms like the decision tree algorithms are used to correctly identify the gear specifications based on the amount of torque applied from factors like material, hole sizes, and maximum force that can be applied at certain temperatures pertinent to regular working conditions maintaining a factor of safety of 2. The decision trees were able to correlate various gear parameters and material properties with high accuracy to provide the means to select the gears based on three types of torques; from high torques of more than 10,000 N mm to low torques of less than 1000 N mm. © 2021, Springer Nature Singapore Pte Ltd.