Transportation construction projects are often plagued by cost overruns and delays. Applying contingencies and estimating risks at the project level often do not capture the multiple uncertainties in the construction process of large transportation projects. Thus, there is a need for innovative approaches and tools to avoid large construction cost and duration overruns. To counteract such underestimations, a construction model and an uncertainty model are developed. In the construction model, the construction of the four main types of structures in rail lines (tunnels, viaducts, cuts, and embankments) is modeled bottom-up from the single activity to the entire rail line. In the uncertainty model, three sources of uncertainty (variability in the construction process, correlations between the costs of repeated activities, and disruptive events) are modeled jointly at the level of the single activity. In a Monte Carlo simulation environment, these uncertainties are propagated to the total construction cost and duration through the combination of the individual activity costs and durations. The construction and uncertainty models are incorporated in the decision aids for tunneling (DAT), which have been extended beyond tunneling to consider these different structures and uncertainty types. All this was applied in the Portuguese high-speed rail project, in which historical data and expert estimations were used to model the cost and duration uncertainty. This application allowed validation of the model and then illustration of a variety of effects: the three sources of uncertainty produce different cost and duration impacts depending on the type of structure, suggesting structure-specific mitigation measures. Most importantly, their cumulative impact causes significant increases in construction cost and duration of the modeled rail line compared with the deterministic estimates: specifically, 58% in the construction cost of tunnels, and 94% in the construction duration of cuts and embankments. The proposed construction and uncertainty models contribute and advance the body of knowledge: For the first time, variability, correlations, and disruptive events are quantitatively modeled in one simulation environment, and the impact of these uncertainty sources can be assessed jointly and compared. The proposed models also significantly contribute to practice by providing transportation agencies with a modeling tool to tackle cost and duration uncertainty in the construction of rail lines and other linear or networked infrastructure projects.