Abstract: When testosterone is detected in food at levels approaching the critical limit of 2.0 μg/kg, uncertainty assessment becomes essential to improve the accuracy of detection results and provide data support for determining food compliance. Based on the Ministry of Agriculture Announcement No. 1031-1-2008 “Detection of 11 Hormone Residues in Animal-Derived Foods by Liquid Chromatography-Tandem Mass Spectrometry”, JJF1059.1-2012 “Evaluation and Expression of Measurement Uncertainty”, and JJF 1135-2005 “Evaluation of Uncertainty in Chemical Analysis”, this study employed ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) with isotope internal standard quantification to establish a mathematical model for analyzing uncertainty sources and evaluating the uncertainty of testosterone determination in goat meat. Results showed that when the measured testosterone content in goat meat was 2.00 μg/kg, the relative combined standard uncertainty was 4.50×10⁻² μg/kg, with an expanded uncertainty of 0.18 μg/kg (confidence level 95%, k=2). The contribution rates of various uncertainty components, in descending order, were: standard solution preparation, sample measurement, standard curve fitting, recovery of spiked blank samples, and sample pretreatment. Evidently, when the detected testosterone content equals the regulatory limit, direct determination of compliance is not possible. More precise pipettes should be used for standard solution preparation, working standard solutions tailored to sample content should be prepared, and parallel measurements should be increased to control the three major uncertainty components: standard solution preparation, standard curve fitting, and sample measurement, thereby ensuring the accuracy and reliability of measurement results.