It has been shown by Tam and his coworkers [7] that the support vector machines (SVM) have excellent capability of handling complicated single-phase heat transfer problems. Therefore, it would be logical to extend the investigation to other flow situations, such as two-phase flow or flow in microtubes. In this study, SVM is used to correlate the two-phase, two-component flow data. Four sets of experimental data (a total of 255 data points) for vertical pipes used in this study were from Kim et al. [3]. They proposed a heat transfer correlation for turbulent gas-liquid flow in vertical pipes with different flow patterns and fluid combinations. Their correlation predicted the experimental data with a deviation range of -64.7% and 39.6%. Majority of the experimental data (245 data points or 96% of the data) were predicted within the ±30% range. A new correlation using SVM is developed in this study. The new correlation outperforms the traditional least-squares correlation and predicts the experimental data within the ±15% range.