An in vitro steroidogenesis assay using H295R human adenocarcinoma cells is a useful tool for the fast identification of compounds that affect the production of testosterone and 17β-estradiol. Selective and sensitive hormone measurement by liquid chromatography-tandem mass spectrometry (LC-MS/MS) can make this assay more reliable. Therefore, in the present study, a sensitive and selective method for the quantification of testosterone and 17β-estradiol in the H295R steroidogenesis assay was developed and fully validated using LC-MS/MS coupled with an online sample enrichment technique. To prove its usefulness, the method developed was applied to investigate the effect of sildenafil on steroidogenesis. Cell medium samples were diluted and prepared using solid-phase extraction. The samples were prepared on ice and were not kept for more than 30 min to prevent degradation of hormones. The extracts were dried, reconstituted, filtered, and analyzed by LC-MS/MS with polarity switching electrospray ionization. The validation results for selectivity, matrix effect, recovery, linearity, precision, and accuracy were satisfactory. The limits of detection for testosterone and 17β-estradiol were 5 and 10 pg/mL, respectively, and the limit of quantification for both testosterone and 17β-estradiol was 10 pg/mL, which was in accordance with the OECD guideline. No degradation was observed under the storage conditions for 7 and 14 days at -80 °C as well as after three freeze-thaw cycles, whereas 17β-estradiol was degraded after 1 h on ice during sample processing. The method developed was successfully used for the investigation of the effect of sildenafil on steroidogenesis. This method can be very useful for the initial selection of drugs with androgenic and/or estrogenic effects for specific purposes, e.g., in the selection of drugs that are used to reverse the effects of chemical castration.