Although most perennial tropical and subtropical grasses are unable to overwinter when introduced into southwestern low-land area of Japan, it seems to be modifiable by the cultivation treatments in the preceding year. Using several tropical and subtropical grass species, three field experiments were carried out to determine the effects of the seeding time, the cutting frequency and the final cutting date in autumn of the seeding year on the winter survival and the productivity in the second year. Comparin g to the treatments of early seeding and three or four times cutting, the late seeding and the two times cutting (clipped both at heading stage) decreased the survival percentage and the spring regrowth yield. But the effect of the final cutting dates on the winter survival showed differences among grass species. It was suggested that those treatments which influenced the growth and simultaneously some plant constituents (mainly TNC) at immediately before the winter affected indirectly the winter survival through those influences. Highly winter hardy species which were measured by field survival percentage were superior in the total yield in the second year, and the treatments which raised winter survival resulted also in this way. Although Buffel grass (Bg) and green panic (Gp) are perennials, these species should be used as annuals for the reason of low chilling resistance before winter. And by rotating with temperate winter annual grasses, these species may be able to result higher year round total production than the perennial or short term single plant grasslands of dallisgrass (Dg), Bambatsi makarikarigrass (Bp) or Kazungula setaria (Ks) .