Steam injection method as a thermal oil recovery technique is a recent remarkable technology for recovering bitumen from oil sands. An interface is formed by contacting the injected steam and the oil sand. Therefore, it is very consequential to study on bitumen-water interfaces since they are expected to play important roles in bitumen recovery from oil sands. A relevant measurement on bitumen-water interfaces would be the dynamic interfacial tension (IFT) of the bitumen-water system at different conditions. As experimental methods usually take a long time and are not economical, the feasibility of using robust tools in the prediction of the dynamic IFT of bitumen-water systems was conducted in this contribution. For this purpose, the least squares support vector machine (LSSVM) model optimized by coupled simulated annealing (CSA) was developed as a powerful method to estimate the dynamic IFT of the bitumen-water system as a function of pressure, temperature, density difference, pH, asphaltene content, bitumen acid index and time. The model results were in an excellent agreement with actual data with a coefficient of determination of 0.99 for the training set and 0.95 for the test set. Furthermore, the proposed model reliability for IFT prediction was investigated by evaluating its results over input variables. The presented LSSVM-CSA model is a very powerful and simple to use approach for predicting the dynamic IFT of a bitumen-water system.