The research demonstrates that high-frequency pulse pneumatic conveying systems are widely used for transporting the lightweight, elastic rod-shaped materials. However, the operational instability that is caused by significant pressure fluctuations, critical component wear, and material coupling frequently leads to counting errors and sudden pipeline blockages. To overcome these challenges, this paper proposed a high-precision intelligent monitoring system and a comprehensive predictive maintenance framework. The sensing unit with a 10 kHz response frequency appears to capture transient waveform features of materials. This unit examines materials moving at high velocities that exceed 15 m/s. Moreover, an adaptive Gaussian clustering algorithm is developed. This algorithm is distinct from high-latency deep learning models. The method resolves the superposition problem. Furthermore, experimental verification demonstrated that adaptive Gaussian clustering algorithm achieves identification accuracy. The accuracy is 97.8% for multi-rod coupling events with computational latency of 0.08 ms. Thus, the algorithm significantly outperforms two-dimensional convolutional neural network and support vector machine. Additionally, a composite health index model is established to quantify system degradation. Field tests indicated that composite health index logic triggers early warnings. Nevertheless, the system maintained long-term counting error below 0.01%.