Different sampling strategies are simulated by changing quadrat size, quadrat shape, sample size and the arrangement of quadrats in a tropical rain forest of Hainan (South China). The simulation uses enumeration data of trees, and derived variables such as species richness, species importance, and species population density, to compare the efficiency of the sampling. The results verify that greater sampling efficiency is to be expected using systematic sampling than random sampling. Quadrat size has substantial influence on parameter estimation, but quadrat shape has negligible effect except when the quadrat is extremely long and narrow.