Acquaintance of the lymph node is very important to evaluate whether the tumor is malignant or benign and therefore to the treatment of the tumors. Lymphography is now considered the gold standard for this purpose. Many computer aided diagnose (CAD) technologies have been developed to help radiologists to diagnose the tumor by the lymphography cases. In this paper, a computer aided diagnose model is constructed by Semi-naive Bayes Classification. The experiments carried out in our laboratory validated the Semi-Naive Bayes Classification on lymphography case. The result of experiments showed that Semi-Naive Bayes Classification could classify lymphography case effectively.