The problem that often occurs in Neural Networks is how to determine the right model, see effectiveness, accuracy, based on parameters, and datasets. Furthermore, various training and testing are needed to determine the best model. In-depth Artificial Neural Network Analysis is required to be analysed and compared with various other classical methods. This study tries to offer the best analytical method, namely Artificial Neural Networks that can be used, especially in the data category. Solution The analytical approach used is to add additional layers to the model until the accuracy is >= 90%. The subject of this research is the health problem of the coastal community, namely the classification of the dominant disease that most suffers from the coastal community. The purpose of this study is to classify using the Protis Neural Network Method approach and compare it with other methods to classify the dominant disease suffered, then measure the level of accuracy of each method and find the best method to classify the dominant disease suffered by coastal communities. The parameters used are region, profession, education, environment, access to health, weather, and dominant disease. The research locations include the regencies of Belawan, Serdang Bedagai, and Central Tapanuli in the Province of North Sumatra, Indonesia. The data used is 100 data. The results showed that the Protis Method approach is known to be able to classify and predict well that the dominant disease suffered by coastal communities is diarrheal disease. From the several methods tested, it is known that the Neural Network Protis is the best method in classifying the dominant disease suffered by the community with an AUC value of 0.9967, a CA value of 0.9504, an F1 value of 0.9505, a Precision value of 0.9523 and a Recall value of 0.9504.