Since about a decade, the modelling of the biological properties of molecules constitutes an important field of research, orienting not only the isolation of biologically active molecules from natural sources, but also the synthesis of active compounds as a potential drug. In this paper, an attempt was made to develop the docking studies of aspirin and a series of one hundred substituted N-arylanthranilic acids with cyclooxygenase protein (PDB-code 2AW1). Molecular docking analysis was carried out using arguslab 4.0.1. The results of the docking software suggested that 56 out 100 molecules have shown best ligand pose energy, the maximal energy is -8.40 kcal/mol and minimal is - 11.18 kcal/mol, among these 56 molecules 20 show a good binding with cyclooxygenase protein, more than aspirin. A Lipinski rule was studied for the best five poses, four of these five molecules satisfy this rule. A computer system PASS (Prediction of Activity Spectra for Substances) was also used to predict the probability of this set of molecules to be anti-inflammatory active/inactive. PASS predict that 73 out of 100 molecules show a good probability of anti-inflammatory activity. All these results lead us to conclude that more than 50% anthranilic acids molecules are suitable for drug treatment of inflammation with less side effects.