Stock co-movement was examined in Finance research but not in the IT research. Previous studies revealed that the co-movement is usually caused by either the determinants of the stocks’ values, habitat movements between stocks, or the change in portfolio composition. Most of the studies used a statistical approach to uncover the co-movement relation between stocks. This paper takes a combination of the statistical approach and the machine learning approach to: (1) prove the existence of stock co-movement; and (2) identify a prediction model that can forecast the stock co-movement. Both supervised and unsupervised methods are used. In this study, the inter-day stock data in the real estate industry were extracted from the Yahoo finance in Hong Kong. After cleaning the data, stocks of the industry were categorized into two groups by its market capitalization. The correlation between the two trading data set is tested. Support Vector Machine (SVM) is used to train the prediction model. The predictive power of the model looks good.