Crystal Structure Prediction (CSP) is used by the pharmaceutical industry to assess whether other polymorphs of active pharmaceutical ingredients (API) might cause problems during manufacturing processes. In the 7th Blind Test of CSP, organized by the Cambridge Crystallographic Data Centre (CCDC), one of the targets (XXX) was to predict the possible stoichiometries of two co−crystals of cannabinol (CBN) and tetramethylpyrazine (TMP). This thesis describes the methodology used for the submission of predicted structures of these co−crystals, concluding that the likely stoichiometries were 1CBN:1TMP and 1CBN:2TMP, as these were more stable than the component structures and had plausible crystal packings. Following submission, this thesis analysed the crystal structures of TMP and have proposed starting points for the crystal structure refinement of a structure (MPYRAZ03) on the Cambridge Structural Database (CSD) that has no atomic coordinates. The CBN search failed to find the Z’=2 experimental crystal structure (CANNOL) that is on the CSD, which has a high energy molecular conformation. This failure was found to be due to the limits on the structure generation program (Sobol sequence and density setting) and was exacerbated by the point charge model failing to model the CANNOL hydrogen bonding adequately. Alternative strategies to find the experimental structure were proposed, but they were deemed too expensive to run a full search. As this thesis was being completed, experimental co−crystal structures were provided by CCDC. After comparing with experimental structures, there was no experimental co−crystal structure in co−crystal CSP searches used in this thesis. This problem was caused by the folded pentane tail instead of the hydroxyl group.