In the search for new co-crystal forms, many studies only consider one method of co-crystallisation which may lead to incorrect results. In this work, we demonstrate the efficiency of applying multiple experimental and virtual screening methods for a more comprehensive search for co-crystals of acetazolamide. A new co-crystal of acetazolamide with 4-aminobenzoic acid ([ACZ + PABA] (1 : 1)) was discovered, although previously, it had been found in the blind spot of the liquid-assisted grinding (LAG) screening method. The new co-crystal was investigated by different analytical techniques, including the powder and single crystal X-ray diffraction, differential scanning calorimetry, dissolution and solubility methods. The specific features of the mechanochemical formation process for [ACZ + PABA] (1 : 1) were studied. It was found that the appearance of the blind spot of the LAG screening method can be caused by a number of reasons; among those are the high sensitivity to the solvent choice and the low rate of the reagent conversion into the reaction product. A comparison of the ACZ co-crystals with 4-aminobenzoic and 4-hydroxybenzoic acids revealed their close resemblance in terms of the packing energy gain and the driving force of co-crystallization. Therefore, the experimental problems in the formation of the [ACZ + PABA] (1 : 1) co-crystal were associated with a number of kinetic reasons, e.g. the high energy barrier of the nucleation process and the low growth rate of the co-crystal. Using the co-crystal screening of acetazolamide as an example, the effectiveness of five different virtual methods for predicting co-crystal formation was assessed. In order to carry out the virtual screening based on the formation thermodynamics of a hypothetical co-crystal, for the first time ever we studied the ACZ sublimation process. Four out of the five virtual screening methods confirm the formation of the new [ACZ + PABA] (1 : 1) co-crystal.