Liquid chromatography-tandem mass spectrometry (LC-MS/MS) shows great promise in clinical application for its high specificity, high sensitivity and wide linear range for the determination of small molecules. However, its application in clinical laboratory is hampered by matrix effect of clinical samples which could greatly affect quantification accuracy and the difficulty to be automated for the traditional sample preparation procedures. Thus, new techniques which could achieve selective enrichment to minimize matrix effect and automatic sample preparation of mass spectrometry are needed. We developed an immunologic mass spectrometry (iMS) method to overcome matrix effect and its clinical application was demonstrated for automatic analysis of testosterone (T), progesterone (P) and estradiol (E2) in human serum simultaneously. Firstly, three monoclonal antibodies were coupled to magnetic beads for selective enrichment of target hormones from serum. The immunomagnetic beads were separated, washed and eluted automatically for LC-MS/MS analysis. Analytical performance of the iMS method was validated and compared with traditional LC-MS/MS and chemiluminescence immunoassay (CLIA). Hormone levels were measured for 160 pregnancy women at different gestational weeks. Results showed that target hormones could be selectively captured with absolute recoveries of 93.9%–110.8 %. Relative responses for high, medium and low concentrations of the hormones between serum and methanol solution were 98.0%–109.7 %, 92.2%–105.3 % and 91.7%–96.0 % for T, P and E2, respectively. Calibration curves prepared in methanol solution, BSA solution and blank serum showed good consistency for the iMS method. The automated iMS method could overcome matrix effect of LC-MS/MS and cross-reaction of CLIA. Matrix effect of the iMS method was negligible as high specificity of target hormone enrichment before LC-MS/MS analysis. Matrix-matched calibration standards were no longer necessary for accurate quantification, which was of great benefit for the clinical application of mass spetrometry.