Accurate prediction of Unconfined Compressive Strength (UCS) for Cement-Treated Base (CTB) is vital in engineering design and construction. Artificial Neural Networks (ANNs) are powerful tools for capturing complex nonlinear relationships and have been used for modelling the UCS of CTB with several input parameters. However, understanding the sensitivity of input parameters in the developed ANN model is crucial for reliable predictions and enhancing model performance. In this study, a comprehensive sensitivity analysis of an ANN model developed for predicting UCS of CTB is presented. The analysis aims to identify the most influential input parameters and their impact on the model's predictive accuracy. A dataset comprising properties like dry density, curing days, cement content, percentage of reclaimed asphalt content, D60, D30, and moisture content, along with corresponding UCS values, was collected for training and testing the ANN model. Initially, a single feed-forward neural network architecture was established and trained using a back-propagation algorithm. The model's performance was evaluated using mean squared error, coefficient of determination, and mean absolute error. The results indicate that dry density, curing days, and cement content significantly influence UCS prediction. The percentage of reclaimed asphalt content and D60 also demonstrate considerable sensitivity, while D30 and moisture content have relatively lower impacts. This sensitivity analysis provides valuable insights into the key parameters affecting UCS prediction using ANN models for CTB. The results can guide future research and assist engineers in selecting the most influential parameters for accurate and efficient UCS estimation in geotechnical engineering applications.