Abstract The National Computer-Based Assessment for SMP level is a quality assessment program for all SMP level schools. This has only been simulated in 2019 and in 2021 this is the first stage of testing. Adaptation to ANBK needs to be done quickly so that the School Quality Score becomes good from time to time and the main goal of the education unit, namely the development of student competence and character, is achieved. Finding a solution to improve the quality of SMP in Sawahlunto City based on ANBK using ANN with Backpropagation Algorithm. The data used in this study was sourced from the Education Office of Sawahlunto City where as many as 11 schools participated in ANBK at the junior high school level. Based on the 2021 ANBK simulation data, the results obtained are above the minimum competency of 11 schools. Furthermore, the data is processed using Matlab software. The processing implementation involves four input variables (reading literacy, numeracy, character survey and learning environment survey). Of the 33 data tested using variations of test data and training data, which are then processed using variations in the learning rate and number of epoch parameters. From the test results obtained the level of accuracy of pattern recognition on the backpropagation method with a learning rate variation of 0.2 and the number of epochs 1000. The results of testing this method are as many as 11 junior high schools that have passed. So that the level of accuracy is 99,9987%. The prediction results of ANBK SMP in Sawahlunto City can already describe the quality of education in SMP in Sawahlunto City. With an achievement level above the district/city average of 36.36%, it can become accurate information to improve the quality of teaching and learning and improve student achievement.   Keywords:   Artificial Neural Networks; Backpropagation; ANBK; Computer-Based National Assessment; Student;                     Education.