The estimation of time delays in construction projects represents a challenging undertaking, frequently constrained by insufficient data, inherent uncertainties, and potential risks. Nevertheless, it remains a crucial element in ensuring the success of a construction project. Marine construction projects represent a highly specialized subcategory of the construction sector, characterized by a considerable degree of risk and significant financial outlays. Despite the extensive application of Machine Learning (ML) techniques across a range of domains, there is a notable absence of studies evaluating their efficacy, particularly in the context of marine construction project assessment. In light of the above, the present study examines the potential of ML techniques for estimating time delays in marine construction projects. A total of 43 factors that affect marine construction projects in terms of time delay were identified and categorized into nine major groups through a detailed analysis of interviews with experts from the marine construction industry. The relative importance index method was employed to ascertain the relative importance of the factors affecting delays. The factors and groups were then ordered according to their level of impact on time delay. Considering the advancements in ML, this study utilizes General Regression Neural Networks (GRNN), Support Vector Machines (SVM), and Tree Boost functionality to estimate the time delay of marine construction projects. To evaluate the predictive capacity of each model, they were assessed using five statistical features and Taylor diagram visualization. With regard to predicting time delay, the overall performance of the GRNN was found to be more accurate than that of the other models, while the SVM model exhibited the least predictive capabilities. The GRNN model was found to be both efficient and precise and, therefore, may serve as a practical tool for predicting the time delay of marine construction projects.