Abstract: Protected by the complex macromolecular structure of kerogen/asphaltene, bound biomarkers show strong resistance to post-depositional alterations such as biodegradation and thermal maturation, effectively overcoming the issues of easy disturbance, contamination, and alteration of free hydrocarbons, and therefore have high application value. How to efficiently and conveniently obtain bound biomarkers has always been a direction geochemists strive to explore. The on-line hydropyrolysis-gas chromatography-mass spectrometry (HyPy-GC-MS) technology uses hydrogen as the reaction gas and can efficiently release bound biomarkers from kerogens with a higher unit yield than conventional pyrolysis techniques like Py-GC-MS. However, the comparability between bound biomarkers released from kerogens and asphaltenes by Py-GC-MS and HyPy-GC-MS and free biomarkers in extracts remains unclear. Therefore, low-maturity samples from three main source rock strata in the Bozhong Sag were selected. A systematic comparison was made between free hydrocarbons in extractable organic matter (EOM) and pyrolytic hydrocarbons released from the corresponding kerogens/asphaltenes using Py-GC-MS and HyPy-GC-MS. The results showed that the proportion of C₂₇ regular steranes in Py-GC-MS products from kerogens differed significantly from those in the free state, with a higher proportion of alkenes. Compared to Py-GC-MS, HyPy-GC-MS significantly increased the yield and proportion of alkanes. Its products released from kerogens and asphaltenes showed distribution characteristics of C₂₇-C₂₉ regular steranes similar to those in EOM, with ratios like TT₂₃/H₃₀ and H₂₉/H₃₀ closer to the free state and exhibiting low-maturity characteristics. Therefore, in low-maturity source rocks, the HyPy-GC-MS technique can effectively inhibit secondary free radical reactions, better preserve the integrity and original geochemical information of steranes and terpanes, making it a convenient method for obtaining reliable bound biomarkers. However, its universality in high-maturity or strongly altered samples requires further verification.