Abstract: The result obtained with point-centered quarter method by 56 sampling points was compared with the results obtained with actual measurement and count-plot method by 78.5 % sampling ratio of area in phytocoenologicai investigation on a tropical mountane rain forest community in Jianfengling region, Hainan Island. The comparison showed that density,mena area at brest height, dominance, number of component species and order of I. V. of component species obtained with the point-centered quarter method were satiafactory for phytocoenological investigation. The advantage in speed with the point-centered quarter method was particularly great, with a saving of 60% or more in term of time spent by the count-plot method. But the determination on number of sampling points for the expenimental stand was not based on the conclusion, in almost all the cases, a sample of sufficient size to give an adequate sample of distances will produce an adequate sample of basal area as well, which was drawn by Cottam and Curtis, the designers of point-centered quarter method. The number of points required by adequate sampling of mean area at brest height was the number of sampling points for the experimental stand of the tropical mountane rain forest community.