In modern industry, electric motors are essential components in a wide range of applications, from manufacturing production to transportation and power generation. These motors are critical in industrial machinery and equipment, and their proper functioning is crucial for maintaining operational efficiency and productivity. However, motors are susceptible to various types of faults that can be costly in terms of downtime, production loss, and repair expenses. Early detection of these faults is essential to prevent unscheduled shutdowns, reduce maintenance costs, and avoid workplace accidents. This paper proposes a low-cost, real-time fault detection system for motors placed on the Non-Drive End based on neural networks, aimed at improving operational efficiency and reducing maintenance costs in the industry.