Orthotropic steel decks can experience fatigue at welded joints, the assessment of which turns out to be a complex task owing to their intricate geometry, the stochastic nature of the primary live load (traffic flow) and the temperature-dependent composite action between the pavement and the steel deck. In recent years, the possibility of monitoring, in addition to traditional inspections, has been put forward as a means of improved assessment. Nevertheless, a rigorous framework to (a) enable the effective use of high amounts of multiple/incomplete data provided by distributed data acquisition systems, (b) improve current monitoring-based assessment methods, and (c) enhance current simulation and data visualization techniques, is still absent. A theoretical framework is presented in which a stress-related performance indicator is estimated through a multiple regression model with hourly pavement temperatures and heavy traffic intensities as independent variables. The proposed performance indicator is proportional to fatigue damage following the principles of the S–N approach and the Miner’s rule. Typical applications of this model include (a) analysis of monitoring outcomes for performance assessment, (b) performance prediction of past/future events and (c) fatigue assessment. To illustrate the proposed approach, model-based performance predictions are benchmarked with real monitoring outcomes from the Great Belt Bridge (in Denmark), and good agreement has been observed. Moreover, model predictions are used to estimate the fatigue life of a monitored welded joint. The new methodology enhances Structural Health Monitoring (SHM) methods for orthotropic decks and provides a framework to integrate and visualize the multiple outcomes produced by modern monitoring systems as a part of the Bridge Management System or to assess the remaining life of structures.