PCB (Printed circuit board) is that the fundamental carrier in electronic devices on which a number of small electronic components are placed. The performance of devices is affected largely by a fault in PCB. To overcome the inefficiency of manual detection, time, and cost, automated optical inspection (AOI) supported machine vision has been widely utilized in the industry. As PCB designs get more and more complicated, the tasks like detection and classification also get more harder than before. For the purpose of solving the above problems, we have proposed a PCB fault detection system using machine learning. The system will automatically detect the PCB with the help of computer vision and compare it with the working PCB image which will be already saved in the database. After comparison, the application will be able to tell whether PCB is working or faulty and likewise generate a report. It will also recommend a solution to faulty PCB.