Pavement maintenance optimization ensures that road networks remain in good condition and minimize the need for costly repairs and reconstruction by strategically allocating resources and scheduling maintenance activities. This paper proposes a network-level maintenance optimization of pavement for National Highways in Tamil Nadu, India. Pavement Condition Index (PCI) is used to analyze the present condition of sections under study. Multiple Linear Regression (MLR) model is used to forecast the pavement deterioration rate for the analysis period. Optimization is done for both non-clustered and clustered models in General Algebraic Modelling System (GAMS) software using Mixed Integer Programming (MIP). The objective function is to maximize the effectiveness of maintenance expressed as benefit area under the deterioration curve. The constraints are related to budget and choice of maintenance. Results of optimization for non-clustered and clustered models were compared in terms of the benefit area.