Abstract: Any inevitable condition that harms a person's heart is referred to as heart disease. Due to the abundance of electronic health data already available, computer systems have entered the field to support diagnosis utilizing machine learning (ML) techniques. This study uses a multi-layered perceptron (MLP) trained by a back-propagation (BP) artificial neural network (ANN) to categorize the data. The MLP is used to identify heart illnesses that contain or do not include heart attacks. ANN can determine the patterns in the data and then create a model by iterating through layers of functions. This investigation employs sigmoid activation functions and runs over a thousand epochs. The best results came from testing various learning rates and neuron number values. The results showed 25 neurons and a 0.25 learning rate with a high accuracy of 80.66%. It is discovered that ANN can be applied to categories of cases of heart disease.