Multi-line charts are crucial data in scientific research, which are often analyzed through clustering techniques to uncover underlying characteristic relationships. However, processing from a single feature perspective can easily overlook the interrelations between waveforms. Multi-view clustering has the potential to compensate for the limitations of single-view clustering. But expanding the number of views can introduce interference among views, leading to a decline in clustering performance. Therefore, stabilizing the performance of multi-view clustering presents a new challenge. Addressing these issues, this study proposes a time-frequency domain differential entropy enhanced multi-line chart clustering (DE-MCC) method. The core idea of the DE-MCC is to enhance the feature perspective by adding time-frequency domain differential entropy to the multiple views formed by different temporal waveforms, thereby improving the model's ability to capture complementary information from multiple views. It also involves weight-controlled fusion of soft clustering vectors obtained from different view combinations to ensure accuracy while stabilizing the clustering results. The proposed method has achieved desirable results on two complex non-stationary multi-line chart datasets: electroencephalogram (EEG) and electronic energy level charts in materials science. Compared to other advanced multi-view clustering methods, the clustering accuracy for EEG data reaches 79.38%, and the standard deviation of multiple independent experimental results decreases by 47.9%, demonstrating the stability of the proposed method.