A coarse-grained (CG) model for the simulation of nanoconfined water between graphene surfaces is developed. For this purpose, mixed-grained simulations are done, in which the two-site water model of Riniker and van Gunsteren [S. Riniker, W. F. van Gunsteren, J. Chem. Phys. 2011, 134, 084110] is simulated between atomistically resolved graphene surfaces. In the developed pure CG model, the two interaction sites of water and a combination of eight carbon atoms in the graphene surface are grouped together to construct water and surface CG beads. The pure CG potentials are constructed by iteratively matching the radial distribution functions and the density profiles of water beads in the pore with the corresponding mixed-grained distributions. The constructed potentials are shown to be pore-size transferable, capable of predicting structural properties of confined water over the whole range of pore sizes, ranging from extremely narrow pores to bulk water. The model is used to simulate a number of nanoconfined systems of a variety of pore sizes at constant temperature, constant parallel component of pressure, and constant surface area of the confining surfaces. The model is shown to predict the layering of water in contact with the surfaces, and the solvation force is in complete agreement with the mixed-grained model. It is shown that water molecules in the pore have smaller parallel diffusion coefficients compared to bulk water. Well-organized layers beside the surfaces are shown to have lower diffusion coefficients than diffuse layers. More information on the dynamics of water in the pore is obtained by calculating the rate of water exchange between slabs parallel to the surfaces. The time scale to achieve equilibrium for this process, depending on the pore width and on the degree of layering of water beside the surfaces, is a few nanoseconds in nanometric pores.