The precision motion platform is a fundamental component of semiconductor equipment, with its motion performance serving as the foundation for the entire system. In engineering applications, the nonlinear characteristics of the moving platform, such as the nonlinear forces introduced by flexible steering, cable forces, and maglev compensation, significantly impact precision motion control. Traditional feedforward control methods, such as the inverse model feedforward approach, require substantial energy to accurately construct the inverse model of the object plant. Meanwhile, the iterative learning feedforward method demands high repeatability of motion conditions. Additionally, adaptive feedforward control can lead to instability in the system’s transient response during parameter adjustments. To address these challenges, this paper proposes an inverse model feedforward controller based on a physically guided radial basis function (RBF) neural network. This approach leverages the RBF neural network’s superior nonlinear function approximation capabilities and employs the gradient descent method for automatic model optimization, thereby significantly reducing the modeling workload. Furthermore, by incorporating prior experience of acceleration feedforward into the RBF neural network inverse model, the proposed method effectively reduces tracking error and enhances the system’s response speed.