In order to diagnose or evaluate sub-health state, we have used data fusion technology and developed a new method which is based on the hidden information among ECG and pulse signals. The characteristic was constructed with pulse power spectrum peak value and pulse transit time (PTT). The classification of health state was realized by using linear discriminant analysis (LDA) method. The results showed that the pulse power spectrum peak value of sub-health state was decreased significantly when compared with health state [(42.3843 +/- 3.7116) vs. (37.6022 +/- 4.4468), P < 0.01], and the pulse transit time was also decreased significantly [(326.5801 +/- 36.4035) vs. (259.1023 +/- 67.3719), P < 0.01]. The classification accurate rate can reach 90%.