The importance of rubber wood as the timber of the present era is on the increase and therefore accurate estimation of wood available from a large population of rubber trees in a plantation has also gained much practical significance. Sampling is the ideal approach adopted in forest tree species to estimate timber volume and availability from a population. Three sampling procedures, viz. simple random, systematic and stratified, were adopted for estimating wood volume in a study comprising of mature trees of five clones (RRII 105, RRII 203, RRII 208, GT 1 and RRIM 600) in five locations (on-farm trials) within Kerala. Stratification was carried out based on the stable variable, the girth at the thick end of trunk. The sampling intensity required to estimate the population mean within 10% error at 95% confidence level was determined for each sampling procedure. The average sampling intensity was found to be 10, 15 and 21 %, respectively for stratified, systematic and simple random sampling procedures. Error percentage was also estimated for these three methods, i.e. stratified (3.5), systematic (7.0) and simple random (9.0). Since the criterion for judging a sampling procedure is minimum per cent error at a minimum sampling size, stratified sampling was found to be ideal for rubber wood estimation.