Based on the repeatability and intermediate precision of time differences in the standard, through inspection of the data of repeatability and trueness of measurements within the time intervals, repeatability between time intervals, general precision between time intervals and trueness of grand mean of long-term stability test. The long-term stability of analytical instruments can be systematically evaluated according to the standard, the essence of which is to monitor the precision and accuracy of analysis results. However, a large number of test methods in the laboratory that have not been upgraded to standards, the so-called non-standard methods, cannot be directly evaluated for long-term stability according to the above methods due to the lack of repeatability and laboratory reproducibility data. Fortunately, the stimulated repeatability and within-laboratory reproducibility limits were obtained for non-standard test methods after correcting the instrument between different time intervals. Based on the stimulated repeatability limit and within-laboratory reproducibility limit, the long-term stability of the corresponding analytical instrument method can be evaluated. This paper uses the non-standard method for determining 18 impurity elements in pure nickel by glow discharge mass spectrometry (GD-MS) as an example. A set of comparative experiments are designed to obtain the simulation reproducibility limit and simulation reproducibility limit data. The long-term stability time for 18 impurity elements, such as B, Mg, Al and Si in pure nickel samples was evaluated by χ2 statistics after being tested by GD-MS. The results showed that the long-term stability was different for different elements when determined under the same condition. The results met the statistical requirements for most elements within 3 hours. Among these, the long-term stability interval can be up to 6 hours or even 12 hours for elements P, As, V, Sb and Pb. The reliable results can be obtained within that time interval without instrument correction. The evaluation results are consistent with the laboratory experience, indicating that the systematic measurement and characterization method proposed in this experiment objectively reflects the long-term stability of the GD-MS objectively. This method can also be used to evaluate the long-term stability of other non-standard methods, which has significant practical guidance for the quality control for the laboratory.