For several decades, much attention has been paid to thylakoid membranes (TMs) as photocatalysts for converting solar light to electricity. Despite extensive research, current technology provides only limited photocurrents. Here, a novel method based on TM-composite material was developed for achieving high photocurrent. When a thin film composed of TMs, osmium redox polymer (Os-RP), and indium tin oxide nanoparticles (ITOnp) was formed on a porous graphite surface, appreciable photocurrent as high as 0.5 mA cm-2 was achieved at 0.4 V vs. Ag/AgCl. Each component plays its own role in transferring electrons from TMs to the anode, resulting in sharp drop in photocurrent with missing any component. Optimization between these three components showed 1 : 0.5 : 30 (TM/Os-RP/ITOnp) was the best ratio. Action spectra confirmed that TMs was the origin of photocurrent. It was inferred from blocking experiments using 3-(3,4-dichlorophenyl)-1,1-dimethylurea as an inhibitor that about 41 % of photocurrent was transferred from QA in photosystem II to the electrode via Os-RP and ITOnp. Quantum efficiencies at 430 and 660 nm were 12.2 and 18.5 %, respectively. Turnover frequency for water oxidation depended upon the amount of the composite. A complete cell with Pt/C cathode produced Pmax of 122 μW cm-2 at 758 μA cm-2 under one sun illumination, which is the highest power density to our knowledge. This study opened a possibility of using TMs as photocatalysts for solar energy conversion.