The most common kind of arthritis is called osteoarthritis (OA), which is caused by the deterioration of articular cartilage, which is essential for cushioning impact in our joints. The osteoarthritis of the knee is the condition we will be focusing on today. Because MRI, or magnetic resonance imaging, can distinguish between bone and cartilage, it is becoming a commonly used diagnostic tool for evaluating the course of osteoarthritis. In the past, doctors had to manually review MRI pictures in order to interpret them. This was a laborious and inconsistent procedure. As a result, automated classifiers become essential for expediting the categorisation procedure. In this work, we utilise a neural network approach optimised by evolutionary algorithms to classify knee osteoarthritis. This classifier consists of four stages: Discrete Wavelet Transform (DWT) is used to extract features; Genetic Algorithm (GA) is used for the neural network’s development, followed by evaluation, and improvement. Considering an identification precision of 97.31% during testing and 98.0% during training, our method produces encouraging results.