Weak information measurement between the spectra is one of the toughest current research issues in the hyperspectral remote sensing domainTraditional measuring methods are difficult to distinguish the weak information differences. The experiment on the lead(Pb) pollution was designed based on its different concentrations, meanwhile, the hyperspectral reflectance, chlorophyll and lead ion(Pb2+) contents of corn leaves stressed by different Pb2+ concentrations were measured. However, it is difficult to distinguish the differences on weak information between the spectra and the pollution levels of corn leaves stressed by different Pb2+ concentrations because the spectral correlation coefficients have reached 0.999 according to the measured results. Due to this fact, a novel spectral similarity measuring method that is the derivative spectral angle tangent (DSAT) model, was put forward based on the spectral derivative processing, tangent function enhancement, spectral angle measurement, piecewise spectral detection and so on. In order to verify the feasibility and effectiveness of DSAT in distinguishing the differences of the similar spectra that their correlation coefficients reach 0.99, the DSAT was used to measure the weak information differences between the spectra of corn leaves stressed by different Pb2+ concentrations by the ways on detecting the whole waveforms and the sub-interval waveforms of corn leaf spectra. The experimental results showed that the relative chlorophyll concentration and Pb2+ contents of corn leaves were significantly correlated with the waveform difference information. It also proves that the DSAT model has better practicability and superiority in distinguishing the difference between the high similarity spectra.