Pavement networks are among the most valuable highway assets for a nation asbillions of dollars have been invested every year in construction, maintenance, and rehabilitation. These networks undergo deterioration over time due to traffic loading, material characteristics, and environmental factors. Various prediction models were developed to predict pavement performance for several purposes, such as preparing an asset management plan, budget, and investment strategy. However, limited studies were found that were conducted on developing a probabilistic deterioration model for composite pavement networks. Also, most of the pavement management system’s prediction models did not integrate flooding, and very few studies were found in this regard. Therefore, this research aims to develop a cluster-based pavement deterioration model through the Markov Chain and the Monte Carlo simulation analysis for composite pavements and propose a framework for incorporating flooding in the model. To this end, a case study was conducted on 102 pavement sections located in the United States' eastern region. For this purpose, the roughness, traffic loading, temperature, and precipitation characteristics from 2015 to 2019 was collected from the LTPP database. These pavement sections are grouped into three different clusters using the K-means clustering algorithm. Then, with the application of Markov chain analysis and Monte Carlo simulation, the pavement deterioration model for each cluster was developed. This deterioration model is utilized to predict the family-based deterioration trend. The proposed framework for incorporating flooding is utilized to predict the pre-and-post flood IRI values of flood-affected pavement sections.