Crystal packing plays a central role in determining the properties of energetic materials, as the arrangement of molecules in the solid state controls crystal density, detonation performance, and mechanical sensitivity. Crystal structure prediction (CSP) provides a way to assess candidate energetic materials before synthesis by mapping their crystal energy landscapes and identifying low-energy polymorphs. In this work, CSP was used to investigate three borazine-based candidate molecular explosives. A hierarchical workflow was employed, combining large-scale force-field searches across experimentally common space groups with dispersion-corrected density-functional theory to refine lattice energies and optimize promising crystal structures. The resulting crystal energy landscapes predict that one candidate, bearing azido substituents on the heteroaromatic ring, should have a density comparable to that of established CHNO-based explosives, making it a promising target for future synthesis. Overall, the results demonstrate how CSP can focus experimental efforts on the most viable candidates, underscoring the promise of computational screening in the design of new energetic materials.