<p>Predicting pavement maintenance cost is an important factor for planning the budgetary requirements of any road agency at the network level. Pavement condition prediction models play a vital role in predicting pavement maintenance costs. Road agencies in developing countries like India still need pavement condition prediction models. The present study proposes an approach to help road agencies predict pavement conditions and annual maintenance costs. The proposed method was developed on rural roads of Shimoga district, Karnataka, India. The pavement condition data were collected using the Online Management Monitoring and Accounting System (OMMAS) database and by field visual inspection survey. A homogeneous Markov model was developed to predict the future pavement condition and estimate the rural road network's annual pavement maintenance cost. The study results indicated that 59% of the road network would come to a reconstruction state at a 10-year duty cycle if no maintenance is provided. Consequently, the annual maintenance cost predicted for 2031 was Indian Rupee (INR) 751.6 million, with an increase of 17% on average with each duty cycle. The model was validated along with the sensitivity analysis. The sensitivity analysis indicated improved pavement performance reduces maintenance costs and vice versa. The validation of the model was reliable, with a Pearson’s correlation R value of 0.92 and R square value of 0.86 at 95% confidence level. Hence, with this proposed approach, road agencies can predict the annual pavement maintenance cost so that they can plan their budget accordingly for an effective maintenance strategy at a network level.</p>