Selective Electron Beam Melting (SEBM) refers to an additive manufacturing process of building near net shaped components layer wise by iteratively melting of metal powder using an electron beam. Due to the preheating of the material and very high scan velocities of the electron beam, the process allows large processing windows defined by different process strategies. This variety enables the adjustment of part properties by, e.g., targeting certain microstructures. However, this variety has its disadvantage in the need for costly trial and error experiments. Therefore, numerical predictions are inevitable in locating promising process parameter combinations.We present the weak coupling of a Finite-Element (FE) model with a cellular automaton (CA) model to predict the microstructure evolution during SEBM. We use an FE model for the computation of the heat input, since the temperature history mainly influences the subsequent microstructure evolution. The FE model provides precomputed temperature fields for our sophisticated CA-based crystal growth model, which successfully predicts bulk material microstructures. The compound is validated by accurately reproducing the microstructure of an additively build cylinder from CMSX-4 powder, a nickel based superalloy. Special emphasize is laid on the accurate prediction of the microstructure both in the shell as well as the core of the cylinder.