The Internet has provided remarkable advantages for data communication between computers. However, its complexity and openness as well as rapidly evolving technologies have posed computer networks to a major challenge in terms of protecting the security and privacy of information as it is transmitted from one place to another. Therefore, it is critical to make computing safe and secure from malicious hackers and viruses. The current existing methods suffer from low accuracy and system robustness. To overcome such limitations, this paper proposes a multi-expert classification framework for detecting different types of network anomalies. Specifically, different types of intrusions will be detected with different strategies, including different encoding schemes, attribute selections and learning algorithms. The Knowledge Discovery and Data Mining (KDD-99) dataset is used as a benchmark to compare this method with other existing techniques. It is empirically shown that the proposed design outperforms other state-of-the-art learning methods in terms of learning bias and generalization variance.