Recently the machine learning-based intrusion detection approaches have been subjected to extensive researches because they can detect both misuse and anomaly. One way to describe anomalies is by saying that anomalies are not concentrated. It leads to the problem of finding level sets for the data generating density. This learning problem may be converted as a binary classification problem. In this paper, we propose a new method to design RBF classifier based on multiple granularities immune network, and apply this algorithm to detection the data density level set. Experimental results on the real network data set showed that the new classifier has higher detection rate and lower false positive rate than traditional RBF classifier.