An accident of Fukushima Dai-ichi Nuclear Power Plant (FNPP1) in March 2011 resulted in a large amount release of radiocesium (134Cs and 137Cs) into the North Pacific Ocean. Radiocesium deposited on and discharged directly into coastal area of Japan in the north of the Kuroshio Front had been transported eastward in surface layer and reached to the North American Continent by 2015 and early 2016 (Smith et al., 2017). On the other hand, radiocesium deposited on the south of the front, namely the western subtropical gyre, had been transported southward to 15ºN through subsurface layer due to subduction of the subtropical mode water (STMW) by the end of 2014 (Kumamoto et al., 2017). However, time evolution of Fukushima-derived radiocesium in the subtropical gyre is still unclear because observational data of radiocesium in subsurface layer are less than those at surface. Here we show our new radiocesium data obtained between 2015 and 2017 in the western subtropical gyre. Seawater samples were collected from surface to 800 m depth during cruises of KS-15-14 (October 2015), KH-16-3 (June 2016), KM16-08 (September 2016), KS16-09 (Nov. 2016), and KM17-01 (Jan. 2017). The seawater sample was acidified using nitric acid and then radiocesium in the seawater was concentrated onto ammonium phosphomolybdate (AMP). Radiocesium in the AMP was measured using gamma-ray detectors. Uncertainty of the radiocesium measurement was estimated to be about 8 %. A meridional transection of 134Cs in 2015/2016/2017 along 142-145ºE between 25ºN and 35ºN was compared to those observed in Jan./Feb. 2012 (Kumamoto et al., 2014), Oct. 2012 (Kaeriyama et al., 2016), and Mar.-Jun. 2014 (Kumamoto et al., 2017) along 142-149ºE. In Jan./Feb. 2012, a high 134Cs plume was lying in subsurface layer around 300 m depth between 25ºN and 35ºN and the highest activity concentration of 134Cs (more than 25 Bq/m3) were observed at 32ºN. In Oct. 2012, the peak of 134Cs subsurface maximum (about 10 Bq/m3) moved southward to the south of 30ºN. In Mar.-Jun. 2014, there was still the subsurface maximum layer around 300 m depth although the peak concentration decreased to 5 Bq/m3, which was observed at 34ºN. These temporal changes indicate that 134Cs subducted to the subsurface layer just after the FNPP1 accident had remained within the western subtropical gyre by the end of 2014. The southward and then northward movements of the 134Cs peak between 2012 and 2014 imply anticyclonic re-circulation of STMW within the gyre. In 2015/2016/2017, the subsurface maximum layer (~3 Bq/m3) was also observed although the peak layer deepened from 300 m to 400 m depth approximately. This deepening is probably derived from vertical diffusion toward deep layer and erosion by newly-subducted STMW in shallow layer. Subsurface maxima of 134Cs (~2 Bq/m3) at the western edge of the gyre (26ºN/128ºE and 24ºN/131ºE) in 2017 suggest that Fukushima-derived radiocesium have been spread in the whole western subtropical gyre. This work was partially supported by Grant-in-Aid for Scientific Research on Innovative Areas, the Ministry of Education, Culture, Sports, Science and Technology Japan (KAKENHI), No. #24110004.