The accurate estimation of the subgrade resilient modulus is a crucial element in designing durable and resilient pavements as it quantifies the soil’s bearing capacity to withstand traffic loads without excessive deformation or failure. Conventional methods and devices that are used to estimate , such as the Falling Weight Deflectometer (FWD) have several limitations due to operational and security constraints when measuring at the network level. To overcome these limitations, Traffic Speed Deflectometer (TSD) has been developed, which can continuously assess the pavement bearing capacity without requiring traffic control. However, processing the massive amounts of data generated by TSD requires specialized processing techniques such as Machine Learning (ML). Therefore, the primary objective of this paper is to develop an efficient and effective model that can accurately estimate by combining two techniques: (TSD + ML). The study uses numerical simulations of deflection velocity slope under TSD as the forward model. The research employs Support Vector Machine (SVM) algorithm for ML to classify and estimate modulus values from a synthetic and controlled dataset. SVM model exhibits robust performance in both classification and regression analysis, whereas, 95% of the discrepancies between estimated and tested values did not exceed 15MPa. Advanced numerical validation, sensitivity analysis and parameter estimation methods were also conducted to optimize the model’s performance. Overall, the approach proposed in this paper has the potential to make a significant impact in the field of subgrade evaluation and can facilitate the development of better road infrastructure.