Efforts to improve the performance and safety of vehicles include placing active sensing components (e.g., embedded microsensors) within tires result in intelligent tires. One application of intelligent tire is tire force estimation based on accelerometers. However, its development is limited due to the difficulty of relating the tire force to kinematical information by model-based theory. In this manuscript, a universal approach to tire forces estimation by the accelerometer-based intelligent tire is formulated and experimentally validated. First, a microelectromechanical system accelerometer-based intelligent tire prototype is established with the function of on-board monitoring of tire forces. Then, a theoretical rolling kinematics model is proposed for illustrating the mechanisms of acceleration fields, resulting from the coupling effect of rigid body motion and elastic deformation. An analytical model is formulated to estimate the vertical force in real time. Furthermore, the beam model is adopted to describe lateral deformations of the tire belt, directly linking lateral acceleration and lateral force. Finally, the lateral force can be estimated by lateral acceleration and vertical force already estimated. Based on a universal analytical model, the lateral force estimation method realizes high accuracy under different circumstances, even with unified coefficients, by clarifying and eliminating the influence of ply steer. A field test and two bench experiments have been conducted to fully validate the developed model. It can be concluded that the theoretical-analysis-based estimation model realizes an encouraging tire force estimation application with an intelligent tire hardware system.