Accurate identification of on-road and roadside objects is crucial for various applications, including autonomous vehicle navigation, urban planning, and infrastructure maintenance. This chapter presents a novel approach leveraging Mobile LiDAR (MLS) data to effectively classify objects within the roadway environment. The proposed methodology leverages a robust preprocessing pipeline, including noise filtering and ground point removal, followed by efficient clustering to segment objects of interest. A key contribution of this work lies in the development of a Random Forest (RF) based object class prediction model. This model is trained on a comprehensive set of fourteen geometric feature descriptors, carefully selected to capture the unique characteristics of various object classes. The model's performance is evaluated on a semiurban dataset, achieving an overall accuracy (OA) of 97.92% and a Kappa coefficient (κ) of 96.98%. The proposed method demonstrates significant potential for applications in autonomous driving, urban planning, and infrastructure management.