The NMR-observable nuclei of the acidic and basic compounds experience pH dependence in chemical shift. This phenomenon can be exploited in NMR titrations to determine p Ka values of compounds, or in pH measurement of solutions using dedicated pH reference compounds. On the other hand, this sensitivity can also cause problems in, for example, metabolomics, where slight changes in pH result in significant difficulties for peak alignment between spectra of set of samples for comparative analysis. In worst case, the pH sensitivity of chemical shifts can prevent unambiguous identification of compounds. Here, we propose an alternative approach for NMR identification of pH-sensitive analytes. The 1H and X (13C, 15N, 31P, ...) chemical shifts in close proximity to the acidic or basic functional group should, when presented as ordered pairs, express piecewise linear correlation with distinct slope, intercept, and range. We have studied the pH dependence of 1H and 31P chemical shifts of the CH3-P moiety in urinary metabolites of nerve agents sarin, soman and VX using 2D 1H-31P fast-HMQC spectroscopy. The 1H and 31P chemical shifts of these chemicals appear in very narrow range, and due to subtle changes in sample pH the identification on either 1H or 31P chemical shift alone is uncertain. However, if the observed 1H and 31P chemical shifts of the CH3-P moiety of individual compounds are presented as ordered pairs, they fall into distinct linear spaces, thus, facilitating identification with high confidence.