[Background] Anaerobic hydrogen-producing granular sludge has more biomass, better settleability and hydrogen production efficiency than flocculent sludge. Proteomic study on hydrogen-producing granular sludge is helpful to reveal the molecular mechanism of microbial metabolic changes in granular sludge, which is beneficial to optimize the anaerobic metabolism process. However, sample preparation method for its proteomic analysis has not been reported. The gram positive bacterium Ethanoligenens harbinense YUAN-3 has been the only reported auto-aggregative hydrogen-producing strain, which can form hydrogen-producing granular in batch and continuous cultures. The genome of strain YUAN-3 has been completely sequenced. Therefore, granules which were formed by strain YUAN-3 were used as model samples in the following experiments. [Objective] In order to facilitate proteomic research on hydrogen-producing granular sludge, protein sample preparation method for proteomic analysis was optimized. [Methods] Protein yields from liquid nitrogen grinding, sonication, and homogenization on was investigated. Protein quality of samples prepared by trichloroacetic acid (TCA)-acetone precipitation method and phenol extraction method was determined by Two-dimensional gel electrophoresis. The proteins were further labeled with iTRAQ (Isobaric tags for relative and absolute quantification) or TMT (Tandem mass tag) reagent and analyzed by using mass spectrometry. [Results] The protein yield from liquid nitrogen grinding, sonication, and homogenization were 2-fold, 3.9-fold, and 5.2-fold higher than the control respectively. Compared with TCA-acetone precipitation method, the protein spots, especially including basic proteins and low molecular weight proteins, significantly increased and uniformly distributed in the phenol extraction sample. 1 797 and 1 644 proteins were identified in the iTRAQ and TMT labeling samples respectively compared to 2 701 predicted proteins from E. harbinense YUAN-3. The distribution of identified proteins across different ranges of molecular weight, isoelectric point and subcellular localization was similar to the predicted proteome of E. harbinense YUAN-3 based on its genome with some exceptions. [Conclusion] Combination of homogenization and phenol extraction method works the best in sample preparation for proteomic analysis of hydrogen-producing granular sludge, which contributes to obtain high quality proteomic data.