The determination of trace elements in Geological Samples by Atomic Absorption Spectrometry is simple, rapid, accurate and economical, which has been widely used in geological laboratories. However, the complex pre-processing process and testing process will inevitably introduce uncertainty. According to the general requirements for the competence of inspection and calibration laboratories, the uncertainty of measurement results should be properly evaluated. In this study, the concentration of Cu, Pb, and Zn in the national standard rock sample and a core sample from the Qujia gold mine in Jiaodong were determined by Atomic Absorption Spectrometry after electric heating plate digestion. Three times the standard deviation of the blank samples test results was calculated as the detection limits. The results of the standard samples and core samples are by the requirements of DZ/T 0130.3—2006 on the accuracy and precision of the test. The bottom-up method was used to evaluate the uncertainty of results in the laboratory. The sources of measurement uncertainty were determined, including sample weighing, constant sample volume, sample digestion, preparation of standard series, least-square fitting and repeated measurement. The value and expanded uncertainty of six uncertainty components were accurately calculated. Among them, the last four components are the main sources of uncertainty. The results show that the uncertainty of the measurement results ofcopper, lead, and zinc in the standard samplesaresmaller than the uncertainty given in the standard certificate, the concentration of Cu, Pb, and Zn in the core sample is (4.965±0.383), (36.415±2.449), (30.818 0±0.736) μg·g^-1, respectively. The uncertainty of six sources is compared, and some improvements are put forward when the content of Cu, Pb, and Zn in rock samples is measured by this method: adjusting the sampling mass or constant volume to improve the concentration and absorbance of elements in the solution to be measured, adjusting the concentration of standard series to make it close to the concentration of elements in the solution to be measured, increase the measurement times of standard point and the solution to be measured, the pipette with small relative standard uncertainty should be used as much as possible if necessary in order to reduce the measurement uncertainty. Evaluating measurement uncertainty as an effective tool to guide the improvement of analytical methods and test process is of great significance in the accurate determination of trace elements in rock samples.