In this work, crystal structure prediction (CSP) calculations were used to guide the crystallization of three elusive polymorphic forms of meloxicam (MLX): II, III, and V. Despite having been discovered over 20 years ago and described in terms of their powder X-ray diffraction patterns, none of them were reported to be crystallized in a pure form by any researcher other than the authors of the original MLX patent. Our numerous crystallization experiments to repeat the disclosed examples failed and led us to employ CSP calculations to indicate possible crystallization pathways to these forms. The obtained crystal energy landscape featured primarily structures stabilized by NH•••O�S dimers, present also in the most stable, commercially available crystal form of MLX, but several low-energy and low-density structures stabilized by NH•••N dimers were also found. As a result, it was suggested to attempt (1) desolvation experiments which have a capacity to lead to low-density structures, (2) employing crystallization additives with the ability to influence hydrogen bond interactions, and (3) fast solvent removal techniques to promote kinetically stable forms. The performed experiments led to the crystallization of pure forms II, III, and V, in addition to the discovery of new MLX solvates and one cocrystal with pyrazole. Desolvation was found to be the best technique leading to elusive polymorphs. Solid-state NMR experiments performed for the obtained solids proved that in all three forms an NH•••N hydrogen bond motif is present, as indicated by CSP calculations.