Effective clustering of radar emitters can enhance the perception capability of electronic warfare units in the electromagnetic environment. For the problem of radar emitter clustering with an unknown number of classes, this work proposes a method based on threshold segmentation and information fusion for inferring the number of clusters. Firstly, the radar emitter data is preprocessed to obtain the time-frequency information and spectrum information. Then, the threshold segmentation-based key time-frequency information processing algorithm is applied to process the time-frequency information. Next, the obtained key time-frequency information is fused with the spectrum information. Finally, principal component analysis and weighted consensus clustering are used to infer the number of clusters for radar emitters. The proposed method can accurately infer the number of clusters when the signal-to-noise ratio is not less than −4 dB. Experimental results demonstrate that the proposed key time-frequency information processing algorithm and information fusion algorithm can effectively improve the performance of the proposed method, and the performance of the proposed method is superior to traditional methods.