The photocatalytic reduction of CO2 to valuable chemicals and fuels is an efficient approach to control the ever-rising CO2 level in the atmosphere. The present paper describes a significant improvement in photoreduction of carbon dioxide (CO2) using sulfur (S) doped titania (S-TiO2) nanoparticles as a photocatalyst under UV-A and visible light irradiation. The sulfur doping was done by following a simple sonothermal method, and a series of photocatalysts were synthesized with the varied amount of S doping. Various characterization techniques were employed for the photocatalysts such as XRD, surface area, UV-Visible, SEM, TEM, and XPS. The XPS reveals that S is predominantly present as S4+ in S-TiO2. The electronic structure for S-TiO2 anatase was calculated with the Vienna ab initio simulation package (VASP) code in the framework of spin-polarized density functional theory. Additional states closer to the valence band are produced inside the band gap as a result of doping. In situ reductive reaction conditions can partially reduce the catalyst, and results in the shift of Fermi level into the conduction band. It is suggested that S-doping increases catalyst surface conductivity, improves the charge transfer rate and the rate of photocatalytic reactions. The prepared series of catalysts have shown excellent activity under UV-A and visible light for photocatalytic reduction of CO2. The effect of the different base including K2CO3, Na2CO3, NaOH and KOH; catalyst amount; sulfur doping amount; and light wavelength were monitored. Methane, ethylene, propylene, and propane were observed as reaction products. In 24 h, S-TiO2 exhibited the highest photoactivity in KOH aqueous solution with a maximum yield of 6.25 mu mol g(-1) methane, 2.74 mu mol g(-1) of ethylene, 0.074 mu mol g(-1) of propylene and 0.030 mu mol g(-1) of propane under UV-A irradiation. The catalysts were active in visible light and able to generate methane and methanol in acetonitrile-H2O mixture with/without TEOA as sacrificial donor producing 846.5 mu mol g(-1) of methane and 4030 mu mol g(-1) of methanol for the former and 167.6 mu mol g(-1) of methane and 12828.4 mu mol g(-1) of methanol for the latter case. An estimate demonstrates that mass transfer does not limit the CO2 reaction. (C) 2018 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.