According to the procedure for uncertainty assessment described in JJF 1059.1—2012 “Evaluation of uncertainty in chemical analysis measurement” and CNAS—GL06—2019 “Guidance on evaluating the uncertainty in chemical analysis”, a mathematical model was established, and an analytical method for the determination of uncertainty of morphine, codeine, papaverine, narcotine and thebaine in chafing dish food by LC-MS/MS was established. The results showed that the synthesis of relative standard uncertainty for papaverine, narcotine, thebaine, morphine and codeine were 5.1×10^-2, 5.0×10^-2, 5.8×10^-2, 8.4×10^-2, and 8.5×10^-2, respectively; the main factors affecting the uncertainty of measurement results included the sample recovery rate, the repeatability of parallel experiments, sample extraction, calibration curve fitting, and the configuration of standard working solution. From the uncertainty component of calibration curve fitting, it was revealed that morphine and codeine determined by the internal standard method were obviously superior to the other three substances measured by the external standard method. Consequently, using the internal standard method could reduce the uncertainty caused by different sample matrices.