In many parts of the world, dams, hydroelectric power plants, tunnels, and highways have to be built in the regions where the weak rocks are dominant. In the solution of the above-mentioned geotechnical design problems, the elasticity modulus is an important parameter to be found. However, it is quite difficult to find the elastic modulus (E) values of highly porous low-strength rocks with standard laboratory tests. Therefore, researchers are trying to predict the elastic modulus of the rocks with new soft computing methods that use some simple inputs. In this study, the elastic modulus of the low-strength rocks was estimated by the least square support vector machine (LS-SVM) method. The properties of weak volcanic rock samples in different strength, welding degree, and geologies were analyzed. In all the prediction models, the tangent modulus of elasticity is used as a dependent variable, while the seismic P wave velocity(V-t), dry unit weight (a(d)), axial point load strength (Is(50a)), diametral point load strength (Is(50d)), and Brazilian indirect tensile strength (sigma(t)) parameters were used as independent variables. It was found that the LS-SVM method is a more effective tool compared with the multiple regression model in terms of prediction performance. Thus, the results show that if the elastic modulus of the rocks cannot be found directly by the experiment or test findings are suspected, the LS-SVM method may be a good alternative way to estimate the elastic modulus of the low-strength rocks.