Bridges are considered one of the most important infrastructure systems. Frequent assessments of the conditions of the various bridges' structural parts are crucial to reflect the overall deterioration conditions of bridges. Compared to other structural components, the decks of bridges are more susceptible to harsh deteriorations because they are vulnerable to severe conditions such as: enormous traffic loads and varying temperatures. Transportation agencies experience many challenges in devising methods to predict the deterioration conditions of bridges in a precise manner. Previous research works tried to estimate the deck conditions based on a restricted set of data while other studies incorporated a unique modeling technique with relatively low prediction accuracy. Therefore, existing literature has not yet provided reliable models. To this end, this paper tackles this critical knowledge gap using a multi-step methodology. First, the paper identified the key parameters affecting the conditions of bridge decks. Second, three data mining models were developed for the prediction of deck conditions using artificial neural networks, discriminant analysis, and multiple regression. Third, a comparison between the presented frameworks is performed to select the ultimate predictive model with the highest prediction accuracy. The end result is a model that forecasts and assesses the bridge decks' conditions with a good prediction accuracy based on 22 identified variables. This minimizes efforts, reduces time, and cuts costs related to the site inspection of bridge decks.