In order to develop efficient brake and chassis controls, it is often necessary to build comprehensive models of the entire vehicle. This can be challenging during early project phases because there are limited data available. This research introduces a novel hybrid modeling strategy that blends fundamental physics with machine learning (ML) techniques to estimate key vehicle parameters—including inertia, suspension stiffness, and suspension damping—based on easily measurable inputs such as track width, wheelbase, vehicle mass, and center of gravity. Even with relatively small datasets the metamodels developed in this work still show high accuracy and flexibility. A user-friendly interface was created to simplify the estimation of parameters, make predictions, and allow for easy retraining. The robustness analyses and dynamic simulator validations show a strong correlation between the estimated parameters and real-world data. The estimated parameters can be then used in full vehicle simulation models, substituting basic parameters in available reference car models. While recognizing limitations associated with small dataset sizes and the absence of tire-specific parameters—addressable by utilizing existing tire models from similarly sized wheels—the proposed method effectively balances predictive reliability and computational simplicity. This approach enhances the early-phase evaluation of vehicle performance and provides a bridge between simplified conceptual models and complex simulations.