In this paper, we use the Galerkin weak form to construct a structure-preserving scheme for Klein-Gordon-Schrödinger equation and analyze its conservative and convergent properties. We first discretize the underlying equation in space direction via a selected finite element method, and the Hamiltonian partial differential equation can be casted into Hamiltonian ordinary differential equations based on the weak form of the system afterwards. Then, the resulted ordinary differential equations are solved by the symmetric discrete gradient method, which yields a charge-preserving and energy-preserving scheme. Moreover, the numerical solution of the proposed scheme is proved to be bounded in the discrete L ∞ norm and convergent with the convergence order of (h 2 + 2 ) in the discrete L 2 norm without any grid ratio restrictions, where h and are space and time step, respectively. Numerical experiments conducted last to verify the theoretical analysis. KEYWORDS finite element method, Klein-Gordon-Schrödinger equation, structure-preserving algorithm, symmetric discrete gradient method MSC CLASSIFICATION 37K05; 37K40; 65M12; 65M60; 74S05 Math Meth Appl Sci. 2020;43:6011-6030. wileyonlinelibrary.com/journal/mmaAdding Equations (33) and (36) yields Equation (28), ie, the energy conservation law.Theorem 3.4. The solution of the proposed SDG Galerkin structure-preserving scheme (Equations 23-26) satisfiesProof. According to Equations (27) to (28), we have). Then we have the L 2 error estimates as follows:Proof. According to the Equations (24) and (26), we eliminate the intermediate variable v and getwhereNote that we consider the case that the exact solution is bounded in the discrete L ∞ norm, ie, ||p n || ∞ ≤ C, ||q n || ∞ ≤ C, ||u n || ∞ ≤ C. According to Theorem 3.4, the boundedness of the numerical solution has been obtained, ie,