Deterioration models are required and used in Bridge Management System (BMS) to predict the condition and performance of bridges. Effective maintenance of bridge structure relies on the accuracy of deterioration models used to predict bridge performance. Markov Model is a deterioration forecasting model that is widely used in BMS. However, research showed that Markov Model has many shortcomings. This research provides a review of bridge deterioration modeling with emphasis on accuracy improvement of the generated transition probability matrix used in the model. Dynamic Bayesian Network (DBN) technique is utilized to predict the future conditions of bridge decks. Variables such as, factors affecting deterioration process and inspection measurements from Non-Destructive Evaluation (NDE) methods are incorporated to increase the accuracy of the developed deterioration model. The impact of these factors is extracted from the literature and the DBN model is built using a numerical example. Measurements of NDE for years 2008 and 2013 for a case of a bridge deck are used to apply the model. The developed method is expected to improve current practice in forecasting bridge deck deterioration and in estimating the frequency of inspection.