The radio galaxy is an important population of AGNs to study the evolution of massive galaxies, because the host galaxy of radio galaxies is generally characterized by a large stellar mass and old stellar population. The radio galaxy is interesting also for understanding the final phase of the AGN evolution, since radio galaxies are characterized by a relatively high mass of SMBHs and a low Eddington ratio. Especially, radio galaxies showing a very large radio loudness are interesting, since low-z radio galaxies showing a larger radio loudness tend to have a larger M_BH and a lower Eddington ratio. However the number of radio galaxies discovered so far is quite small at high-z. This naturally leads to our interests to study high-z radio galaxies with a very large R, for exploring the early completion of the galaxy-SMBH coevolution in the early Universe. Motivated by this, we focus on the FIRST radio survey because the 70% of FIRST sources are not detected in the SDSS due to too shallow optical images. Therefore, we started a new search for radio galaxies with HSC-FIRST matched catalog made by using the HSC-Wide data which is much more sensitive than SDSS. When we constructed the matched catalog, we adopted a simple positional matching where we set the search radius to 1 arcsecond to achieve high completeness and low contamination. As the results, we found a number of radio galaxies with the radio loudness (in observed frame; R_obs = F_radio / F_opt) higher than 4.5. In addition, there are some radio galaxies whose photometric redshifts are higher than 3. In this talk, we report the current status of our study.