<![CDATA[Serious problems related to waste in urban areas, including Magelang City, have led to deepening problems. Rapid population growth and limited land at Banyuurip landfill make it difficult for the government to manage waste and potentially cause negative impacts on the surrounding environment. In an effort to overcome this problem, it is necessary to predict the amount of waste received at Banyuurip landfill from each village in Magelang city every day. The method used is Backpropagation Neural Network with five steps such as data collection, data preparation, data pre-processing, prediction modeling and model evaluation, with the results showing that the Backpropagation Neural Network method, using parameters 30-7-1 and number of epochs 1000, produces the best Mean Squared Error (MSE) value of 0.00013 in Potrobangsan Village. The importance of data normalization in pre-processing is also emphasized, because it can minimize errors and improve prediction accuracy. In this study, data normalization was carried out using an equation, with minimum and maximum values of 0.610 and 4.600 respectively, which were obtained from actual values.The study also determined the percentage of data division, with 15% for test data (361 data) and 85% for training data (1039 data). The low error rate of the prediction model shows the best performance. The model successfully predicts the daily amount of waste for each day in January 2023 in 17 urban villages in Magelang City, using the Backpropagation ANN method with architecture and models that have gone through training and testing stages.]]>
