Additively manufactured (AM) ceramics present unique challenges and opportunities due to their distinct microstructural characteristics, which may be affected by a wide variety of print parameters and production techniques. This research proposes a method for microstructure-based fracture prediction of AM ceramics, focusing on 94% alumina ceramic printed using stereolithography at Sandia National Laboratories. The methodology for characterizing failure behavior in AM ceramics is based on methodology proposed by Ozaki et al. for conventional ceramics, analyzing observable microstructure and employing finite element analysis (FEA) and a statistical approach to estimate fracture strength. More specifically, the stress state in a component is calculated using FEA, followed by a probabilistic assessment of flaw occurrence based on microstructural data obtained through Electron Backscatter Diffraction (EBSD) or Scanning Electron Microscope (SEM) imaging. This method was adapted to the dense AM alumina ceramics produced at Sandia by considering anisotropy and focusing on grain boundaries rather than internal pores. A mesh refinement study was performed indicating that the FEA mesh size must be appropriately sized in relation to the EBSD or SEM sample size obtained for accurate results. The results of this method were compared against experimental data for two standard ceramic test configurations printed at Sandia. With AM ceramics, it is primarily internal microstructure rather than surface flaws from post-processing that contribute to fracture. Different components printed using exactly the same material, technique, and machine should have very similar internal microstructure. This method thus enables rapid estimation of fracture strength during prototyping, as the same microstructure assessment can be applied to any component for which a stress field can be calculated.