In order to minimise the shortcomings of insufficient inspection records, an integrated and enhanced bridge deterioration method using a combination of state-based and time-based probabilistic techniques has recently been developed. It has demonstrated an improved performance as compared to the standalone probabilistic techniques. Nevertheless certain shortcomings still remain in the integrated method which necessities further improvement. In this study, the core component of the state-based modeling is replaced by an Elman Neural Networks (ENN). The integrated method incorporated with ENN is more effective in predicting long-term bridge performance as compared to the typical deterministic deterioration modeling techniques. As part of a comprehensive case study program, this paper presents the deterioration prediction of 35 bridge elements with material types of cast-in-situ Concrete (C) and Precast concrete (P). These elements are selected from 86 bridges (totaling 1,855 inspection records). The enhanced reliability of the proposed integrated method incorporating ENN is confirmed.