The state of the pavement is a critical component of the transportation infrastructure that affects cost, efficiency, and safety. Machine learning (ML) techniques have attracted a lot of attention recently as a means of improving our understanding and capacity for predicting asphalt pavement behavior. This study provides unsupervised ML models for categorizing the states of asphalt pavement based on their measured features, such as roughness, cracking, rutting, texture depth, potholes, and raveling. Unsupervised clustering algorithms, namely K-means and K-medoids are implemented for clustering pavements based on the conditions relative to deterioration metrics and, allowing the detection of patterns and groupings within vast datasets and the categorization of road segments based on their attributes, even in the absence of labeled data. By choosing variables with significant levels, algorithms were able to cluster the stretches of pavements with almost similar deterioration levels under distinct groups. Optimal number of groups were chosen based on silhouette score, variance within clusters, total variance, and between clusters. The outcomes show that suggested unsupervised ML model effectively partitions different pavement conditions into groups based on their measured parameters, providing insightful information about how asphalt pavements behave. These data can help transportation infrastructure managers make better judgements about maintenance, repairs, and improvements, enhancing cost-effectiveness, safety, and efficiency.