CO2 dissolution in water at different pressure and temperature conditions is one of the most essential issues to CO2 geological sequestration in brine aquifers. In this manuscript, four powerful Machine Learning (ML) techniques including Radial Basis Function Neural Network (RBFNN), Least Square Support Vector Machine (LSSVM), Multilayer Perceptron (MLP), and Gene Expression Programming (GEP), are utilized to generate economic, rapid and reliable models to predict CO2 solubility in water to 3500 bar and 623.15 K. To expand the prediction capability of the ML approaches, their control parameters are optimized by various techniques. Four backpropagation algorithms, are applied in the MLP learning phase, while Differential Evolution (DE), Particle Swarm Optimization (PSO), Firefly Algorithm (FFA), and Genetic Algorithm (GA) are used to optimize the RBFNN and LSSVM control parameters. A wide-ranged database including temperature and pressure as inputs and CO2 solubility in pure water as output, is utilized to...