Naive Bayes (NB) is a simple but powerful tool for data classification. It is widely used in classification due to the simplicity of its structure and its capability to produce surprisingly good results for classification. However, the independence assumption among the features is not practical in real datasets. Attempts have been made to improve the Naive Bayes by introducing links or dependent relationships between the features such as the Tree Augmented Naive Bayes (TAN). In this study, we show the accuracy of a General Bayesian Network (GBN) used with the Hill-Climbing learning method, which does not impose any restrictions on the structure and better represents the dataset. We also show that it gives equivalent performances or even outperforms Naive Bayes and TAN in most of the data classification.