Software-Defined Networking (SDN) is a network architecture that aims at providing high flexibility through the decoupling of the network logic from the forwarding functions. The ease of programmability makes SDN a great platform implementation of various initiatives that involve application deployment, security solutions, and decentralized network management in a multi-tenant data center environment. Although this can introduce many applications in different areas and leads to the high impact on several aspects, security of SDN architecture remains an open question and needs to be revisited based on the new concept of SDN. Current SDN-based attack detection mechanisms have some limitations. In this paper, we investigate two of those limitations: Misbehavior Attack and NewFlow Attack. We propose a secure system that periodically collects network statistics from the forwarding elements and apply Machine Learning (ML) classification algorithms. Our framework ensures that the proposed solution makes the SDN architecture more self-adaptive, and intelligent while reacting to network changes.