<![CDATA[The priority classification of road maintenance is an important issue in regional infrastructure management. This study developed a classification model based on Artificial Neural Network (ANN) to determine the priority of district road maintenance automatically based on actual condition data. The data covered 141 road sections, reduced from 15 to 9 main variables using Principal Component Analysis (PCA), and normalized with the Min-Max Scaler. The ANN model consists of 10 input neurons, 30 hidden neurons, and 5 priority class outputs. The data is divided in a 55-15-35 ratio for training, validation, and testing. The model produces 92% accuracy, 91.7% accuracy, 90.4% recall, and 90.9% F1-score. These findings demonstrate the reliability of ANN in multi-class classifications to support more efficient road maintenance decision-making. The novelty lies in the integration of actual field data, multi-criteria classification, and the application of ANN in the context of complex and underexplored district roads in the literature.]]>
