The characteristics of electrocardiogram (ECG) signal are fuzzy and random, so that they are difficult for automatic analysis and diagnosis. To solve this problem, an uncertainties transformation model-Cloud Model, which is a fusion of qualitative and quantitative information, was tried to use to analyze the ECG signal. The model fusions the characters of fuzzy and random, just suit to the ECG automatic analysis and diagnosis system. Based on the theory of the cloudy transform and comprehensive cloud, the clustering of ECG signal was finished. Further more, the clinic experience of expert was summarized as classification rules based on the theory. The experiment data were from MIT/ BIH database. The experiment results showed more close to those of the expert's analysis. The describing result was more close to those of the more expert's with qualitative and quantitative information. It is well concluded that the method is an effective ECG signal analysis method.