Nitrogen-containing heterocyclic complexes are important parts of many biological molecules including DNA, RNA, proteins and drug molecules, which often have weak intermolecular interactions. Theoretical understanding based on the density functional theory (DFT) of such subtle interactions is of importance but challenging since it is strongly dependent on the choice of the DFT exchange–correlational (XC) functionals. In this study, the performances of 20 DFT XC functionals at the different DFT levels for calculating the intermolecular interactions of nitrogen-containing heterocycles have been evaluated using 23 representative complexes. The benchmark intermolecular interactions are derived from the calculations using the “golden standard” CCSD(T)/CBS method. Our results demonstrate that the XC functionals with the consideration of dispersion forces have the better performance on evaluation of intermolecular interaction. Overall, the functional with the best performance is ωB97M-V for calculating non-bonding interactions of nitrogen-containing heterocyclic complexes, while other XC functionals may outperform it for individual stacking configurations. To reduce the computation cost, the relatively small basis set can be used when the Boys-Bernardi counterpoise corrections are applied.