Since thorium is an abundant fertile material, there is hope for the thorium-cycle fuels for an accelerator driven subcritical system (ADS). The ADS utilizes neutrons, which are generated by high-energy protons of giga-electron-volt-grade, but cross sections for the interaction of high-energy particles are not available for use in current ADS engineering design. In this paper the neutron behavior in the ADS target based on the related experimental data is clarified, and the feasibility of the ADS regarding both the molten salts (Flibe: 7LiF-BeF2-ThF4-233UF4, chloride: NaCl-ThCl4-233UCl4) and oxide ([Th, 233U]O2) fuels is examined. The difference between the experiment and the calculated result at the ADS high-energy region is discussed. In a comparison of the fuels, the time evolution of keff and the beam current in the burning period are calculated. The calculated results suggest that the ADS with solid fuel has better future prospects than that with molten-salt fuels. The ADS with Flibe molten-salt fuel tends to require a high beam current and consequently needs the installation of a metallic spallation target and the continuous removal for fission products and protactinium. In comparison with the Flibe fuel, the ADS with chloride fuel has a flux distribution that is similar to a solid fuel reactor.