Used extensively in structural dynamics analysis, the energy method is advantageous for transforming boundary value problems of differential equations into variational extremum problems. Recent applications include analyzing the wave and vibration characteristics of track structures. However, traditional energy methods for dynamic modeling and analysis of ballastless track structures require obtaining the global stiffness matrix and mass matrix of the entire coupled system, which decreases computational efficiency. The EM-CMS algorithm for frequency domain dynamic response analysis of ballastless track structures has been proposed to address this issue. The core of EM-CMS is the development of model reduction strategies within the framework of the energy method to reduce matrix dimensions and thereby improve computational efficiency. Specifically, the steel spring floating slab track is the focus of this research. Utilizing the energy functional variational method, the modal properties of the rail and floating slab structures are obtained, and truncation is performed. A reduced-order model for the steel spring floating slab track is established by considering the boundary conditions between the rail-floating slab and the floating slab foundation (connected through fastener springs and steel springs, respectively, and considering the springs’ elastic potential energy). Comparatively, the computational efficiency of EM-CMS is approximately ten times greater than that of the energy method. The method’s accuracy is also carefully validated against finite element simulations.