The Summation-by-Parts (SBP) method (Fernández et al., 2014; Kreiss and Scherer, 1974; Mattsson and Nordström, 2004; Svärd and Nordström, 2014) is a spatial discretization technique for partial differential equations (PDEs) that employs the SBP operator to approximate spatial partial derivatives. Using the summation-by-parts formula for this operator, the SBP method mimics the energy method in discrete systems. Furthermore, some SBP operators provide highly accurate approximations of spatial partial derivatives, and their use can be expected to yield high-precision spatial discretization of PDEs. Alternatively, as a structure-preserving numerical method for PDEs, there exists the discrete variational derivative method (DVDM) by Furihata and Matsuo (2010). This method is a structure-preserving numerical method that enables us to systematically derive the structure-preserving scheme. The scheme derived from this method is typically second-order accurate in space. More recently, Umezu et al. (2025) have designed a structure-preserving scheme with high-order spatial accuracy based on DVDM for the Cahn-Hilliard equation with homogeneous Neumann boundary conditions, using the SBP operator (Mattsson and Nordström, 2004) and a specific projection matrix (Olsson, 1995). However, constructing such a projection matrix is difficult for complex boundary conditions, such as dynamic boundary conditions that include the time derivative of an unknown function. Thus, in this study, instead of using a projection matrix, we incorporated a correction term, the SA term, corresponding to the residual of the boundary conditions. This enabled us to design a spatially high-accuracy structure-preserving scheme based on DVDM for the Allen-Cahn equation with dynamic boundary conditions (Gal and Grasselli, 2008). Indeed, through numerical examples, we confirmed that the proposed scheme achieves high spatial accuracy. In this talk, we plan to show you the results obtained.