Abstract: As one of the important topics to universal attention, heavy metal pollution is the significant ecological environment problem around the world. In this paper diagnosis spectral indices for Cd pollution stress in rice were selected, thorough research of hyperspectral response features of physiological ecology parameters change of rice in Cd pollution stress to reveal the remote sensing information mechanism of crops pollution stress. Taking advantage of measured ASD hyperspectral data, we synchronized obtained crop parameters data and Cd content data of farm soil in several time phase of growing season of rice. The response relationships between potential sensitive spectrum remote sensing indices and physiological ecology parameters in Cd pollution stress were analyzed, such as leaf pigment concentration, water content, cell structure and leaf area index(LAI). Then the corresponding diagnosis spectral indices were determined, including Modified Chlorophyll Absorption in Reflectance Index (MCARI), Normalized Difference Water Index (NDWI), Red-Edge Vegetation Stress Index (RVSI) and Ratio vegetation index (RVI). Based on them, multi-level diagnosis spectral indices space was established to represent and distinguish Cd pollution degree in rice. The experiment indicated that this method could diagnose Cd pollution stress in rice effectively, but quantitative estimation precision also waited for enhancing.