The pebble-bed high-temperature reactor is composed of a large number of fuel spheres stacked to form the core and operates with a refueling method that does not require shutdown. To complement this operational characteristic, HTR-PM is equipped with an online burnup measurement system that obtains the activity of nuclides by measuring the gamma spectra. These online measurable isotopes can be used for the establishment of burnup benchmarks, prediction of fissile isotopes, and inversion of burnup power history, and these data have the potential for further exploration. However, when measuring the activity of these isotopes, the calculation of characteristic peak areas, the calculation of detection efficiency curves, and even the intensity of rays and other characteristic parameters of isotopes can all lead to inaccuracies in the measurement results. This paper clarifies the method of gamma spectrum deconvolution for online measurement and the calculation method of characteristic peak areas, as well as the method of internal calibration efficiency, and study their uncertainty analysis methods. Based on the characterized detector and online measurement data of HTR-PM, the study analyzes the measurement errors of online measurable isotopes, including the level of error, the range of error, and the distribution of error. This study not only clearly identified the main sources of uncertainty in online nuclide measurement, guiding the direction for the improvement of measurement techniques, but also provided crucial technical support for the operation and safety management of nuclear reactors. The research findings have significant practical implications for optimizing nuclear fuel management, enhancing the economic viability, and improving the safety of nuclear reactors.