The characterization of Shorea spp. tree species among other forest trees appears relatively complicated. Therefore, certain errors tend to occur during planting stock material collection, particularly at seedling or juvenile stages. This mis-identification could probably be minimized by initial sound identification, although it requires very extensive efforts. As a consequence, precise and rapid identification system is required to differentiate the sample at the seedling phase. The identification process involves usually the use of leaves, in which venation forms a major leaf feature with unique architecture and consistent pattern to segregate Shorea species. However, geometric properties also exist and can be extracted, using a geometric mathematical model. The approach determine the position of venation point by applying the linear coordinate values. This study was aimed at identifying Shorea species, using using random forest classification techniques. In addition, information on leaf venation’s geometric features include the attribute angle, length, distance, scale projection, angle difference, straightness, length ratio, as well as densities of leaf vein, branching and ending points, were necessary. In particular properties, the mean, variance and standard deviation are evaluated. Subsequently, to obtain the most important traits, feature selection was conducted, using Boruta algorithm. The results showed the success of the applied model in classifying Shorea species, by leaf venation feature. Also, optimum accuracy was attained at 91.90%, with cut-off training and testing data of 90:10, by analyzing 1000 single trees. Furthermore, extensive sensitivity and precision values were obtained at 89.95 and 90.66%, respectively. These results clearly indicated a superior performance model.