Spit (Spam over Internet Telephony), known as unsolicited bulk calls sent via VoIP networks, is a major problem that undermines the usability of VoIP. Countermeasures against spit face challenges in identifying and filtering spit in real time. A user behavior based on three parameters (interaction,historical, and social ratio) is used to design an anti-spit technique. The rationale for the technique is that voice spammers behave significantly different from legitimate callers because of their revenue-driven motivations. The scheme uses adaptive training to determine filtering accuracy and estimates the legitimacy of the caller. The ideas can be implemented at the router level for detecting and achieving voice spam control. Compared to existing spit defending techniques, it is simple, fast and effective. Experiments are reported and measure the accuracy based on call intensity, call density, and the size of training data. The proposed scheme inapplicable for detecting and filtering both machine initiated and human-initiated spam calls, better protects VoIP calls against Sybil attacks and spammer behavior changes.