This study evaluates and compares several machine learning methods on the effects of different parameters in lead adsorption capacity. pH, contact time, adsorbent dosage and initial lead concentration were considered as inputs and adsorption capacity was regarded as output. For analysing the input parameters, the response surface methodology was used for experimental designs and the obtained results were utilised here as training sets. Various data mining approaches like support vector machine (SVM), group method of data handling (GMDH), decision tree, random forest, radial basis function (RBF), adaptive neuro-fuzzy inference system (ANFIS) and multilayer perceptron (MLP) neural network were implemented to model the problem and two different optimisation techniques named bat algorithm (BAT) and Grasshopper Optimisation Algorithm (GOA) were employed with MLP and ANFIS model for optimising. By comparing different statistical parameters such as Average Absolute Relative Deviation (AARD), coefficient of determination (R-2), Root Mean Square Error (RMSE) and Standard Deviation (SD), it is found out that SVM method has a considerably better performance relative to other methods for estimating adsorption capacity parameter. Furthermore, coupling of MLP and ANFIS with GOA increases the accuracy of these models.