To address the problem of man-made information scarcity in the machine learning, a novel transfer learning machine based on group probabilities toward the common data, called TGPLM-CD, is proposed. The proposed method is based on the structure risk minimization model, and both knowledges of the source domain and class label group probabilities are considered as well as the common data between the source domain and the target domain in the learning process, which realizes knowledge transfer between the source domain and the target domain. Experiment results on extensive datasets show the effectiveness of the proposed method.