Membrane technology performance can be predicted, analyzed, and designed using machine learning (ML). This involves figuring out the best operating parameters to increase the pretreatment process's effectiveness and stop membrane clogging forming on the membrane layer. Pretreatment was used both before and after reverse osmosis to remove water hardness at the Palm Beach desalination facility on an industrial scale. A laboratory-based experimental methodology was created to assess a number of parameters, including pH, conductivity, salinity, TAC, and TH. Artificial neural networks (ANNs) are the most widely used algorithm for simulating membrane separation-based desalination of saltwater for the removal of water hardness. By contrasting it with the testing dataset, the trained model will be verified. The most practical way to assess the trained model which will be validated by comparing it to the testing dataset. the most usable to evaluate the performance of the developed ANN model as Associated metrics include Mean Absolute Error (MAE = 0.0104), a Root Mean Squar Error (RMSE = 0.028), (prediction average = 0.742) and (R2 = 0.991). The prediction of water hardness after pretreatment process as “ sand filter’ and “Cartridge filters” has an effect on membrane performance as permeate flux variation, treatment cost and successfully analyzed complex large scale water quality evaluation.