Electric Vehicles (EVs) are integral to the pursuit of sustainable urban development because of their substantial potential to mitigate environmental adverse impacts and enhance energy efficiency. The number of EVs is rising rapidly in the modern urban ecosystem. In such a scenario, strategic planning in infrastructure, production, and policy framework formulation critically depends on EV demand’s accurate prediction. This study leverages machine learning (ML) techniques to predict EV demand while incorporating historical EV population data and EV charging stations data within the Washington State of USA. The study proposes a two-phased comprehensive forecasting framework based on Seasonal Autoregressive Integrated Moving Average with eXogenous variables (SARIMAX). Wherein, the first phase deals with data-preprocessing and feature engineering by utilizing information on historical variation in EV counts and EV charging stations (existing as well as planned). The second phase deals with testing of various ML approaches and selection of most accurate algorithm for future EV population forecast. The study also provides regional insights on EV counts in various cities and counties within the targeted state. The study reveals that the proposed simplistic ML technique outperforms others in predicting EV demand. Our findings indicate that these models can effectively account for the dynamic and complex nature of EV adoption trends while simultaneously reducing computational complexities. EV population size is found to still rise rapidly for coming years in the state. The insights derived from this study, formulate a robust foundation for the decision-makers. This paper also underscores the critical role timeseries forecasting in guiding infrastructure development and long-term planning for EV deployment, aligning with sustainability and smart asset management goals.