Ensuring reliable spectrum efficiency in modern wireless communication networks requires robust and automated signal interference management. However, the dynamic and non-uniform nature of wireless environments introduces complex overlapping signals, complicating traditional energy-detection methods. This study evaluates the performance of advanced machine learning and deep learning models for detecting and classifying co-channel and adjacent-channel wireless interference. Through comprehensive experimental testing and simulation, an optimized neural network architecture is identified. Subsequently, the capability of the detection system is assessed under varying signal-to-noise ratios (SNR). The results indicate that while traditional threshold-based methods fail under fluctuating noise floor conditions, the proposed model maintains a detection accuracy above 98% even at low SNR levels down. As a typical example of intelligent spectrum management, this study provides a crucial reference for the optimization of next-generation cognitive radio and 5G/6G wireless network.