Time and cost overruns have become significant obstacles to completing construction projects worldwide. Researchers have long investigated the causes behind these delays and budget overruns to address the problem. This study sheds light on well-established assessment methods designed to predict these overruns. Despite significant efforts to mitigate them, these issues continue to plague the construction industry globally. To gain a deeper understanding, experts were first consulted to review existing knowledge on factors contributing to construction project delays. A hyper-parameter optimized predictive model was developed using artificial neural network (ANN) machine learning algorithms trained with quantitative data. The study, which focused on 191 construction projects completed in Jordan between 2011 and 2021, aimed to delve into the prevalence of time and cost overruns. The findings showed that the model's predictive accuracy for time overruns achieved an R² value of 0.9385 with a Mean Absolute Error (MAE) of 21.7090, while for cost overruns, the R² was 0.938478 with an MAE of 0.057. Notably, the ANN model outperformed existing reference models. The neural network has proven to be a valuable tool in the early design phase, mainly when data for analyzing potential overruns is limited. This approach can yield more accurate predictions and significantly reduce forecast errors.