The strategy for developing human resources for recruitment so that at the same time becoming a reliable workforce as expected is the goal of personnel in certain offices or agencies. This step must be taken by the management of companies or institutions, both public and private, in order to improve human resources (HR). Until now, there has never been any research on the conventional acceptance of prospective employees to test how accurate their performance is. In this study the conventional selection system for prospective employees will be used as a basic concept to find methods for analyzing the performance of prospective employees using computer media with an artificial neural network system approach with the backpropagation method. So that the accuracy of the predictive patterns of prospective new employees is obtained. So that finally the personnel of government and private agencies obtain actual information about the performance of prospective employees who will be accepted as workers. The results of the study using hidden layers and learning constants obtained the fastest convergent value at 3377, and the final results of this study will be published in a national journal accredited SINTA 4 or better.