In India, the transportation industry is the one that uses the most crude oil. Two and three-wheelers are the primary focus for most e-mobility schemes introduced in India. On the other hand, even with a comparatively small vehicle population, commercial vehicles employing diesel engines like buses and trucks account for a major share of nitrogen oxides and particulate matter 2.5 produced by automobiles. Particulate matter 2.5 is considered as most detrimental to human health. Heavy-duty engine manufacturers must meet certification, on-road requirements and prevailing compliance standards set by regulatory organizations to receive engine certification. The vehicle-related emissions are regulated using driving cycles in laboratory measurements. Heavy-duty vehicles play a crucial role in transportation and logistics, representing significant investments for businesses. Ensuring the optimal performance and reliability of these vehicles is essential for operational efficiency and cost-effectiveness. Predictive maintenance, leveraging advanced machine learning techniques has emerged as a promising approach to anticipate and prevent potential failures before they occur. This research focuses on the predictive maintenance of heavy-duty vehicles, specifically targeting the prediction of emitter failures using Convolutional Neural Network (CNN)—Recurrent Neural Network (RNN) as RCNN. Emitters encompass many components such as engines, transmissions, brakes, and cooling systems, whose malfunction can lead to costly downtimes and repairs. Additionally, techniques for handling imbalanced datasets and model evaluation are incorporated to enhance the reliability of the predictive maintenance system. The results of this study should enable proactive maintenance scheduling and resource allocation by giving fleet management and maintenance staff relevant data.