The present project studied the signal drift in liquid chromatography tandem mass spectrometry (LC-MS/MS) and proposed a strategy for compensating such drift. In the study, four 4-component groups were repeatedly run on different LC-MS/MS systems for over 12 h to investigate the dependence of signal drift on time and hardware systems. The 4-component groups each consisted of (1) an analyte, (2) a stable isotope labeled analyte, (3) a compound with similar structure to the analyte, and (4) a compound with dissimilar structure. All of the species showed significant signal drift, generally more than 25% over 12 h. The analyte and its stable isotope labeled analog always have the same drifting pattern including the trends and direction from one LC-MS/MS system to another. Signal drift was also found to be concentration dependent. Our experiments further proved that a conventional stable isotope labeled internal standard in LC-MS/MS quantification would not compensate the variations caused by concentration-dependent signal drift. An ideal internal standard for LC-MS/MS has both identical structure and similar concentration to the analyte. For that, we proposed a new internal standard strategy, pseudo internal standard (Pseudo IS), for LC-MS/MS quantification. Pseudo IS could effectively compensate signal drift in spite of its significant time, system, and concentration dependencies.