To determine organic matter content(OMC)in black soil quickly,hyperspectral reflectance of black soil samples collected from the black soil areas in Heilongjiang Province was measured in laboratory,and characteristics of black soil hy-perspectral reflectance curves in the visible/near infrared bands were studied.Correlation between organic matter content and hy-perspectral reflectance was calculated,and used to ascertain main response spectrum of the organic matter in the black soil.The multivariable linear step regression analysis method was used to build hyperspectral models for prediction of organic matter content in black soil,with spectral reflectance and its derivatives(normalized and differential coefficient spectral form)as independent variables,and ln(OMC)as dependent variable.Root mean squared error(RMSE)was introduced to test the predictability and precision of the models,and coefficient of the determination(R²)was used to evaluate stability of the models.Results demonstrate:(1)620～810 nm bands are the main response spectrum of the organic matter,and the correlation coefficients are all higher than 0.8 in the spectral range,with the maximum up to 0.83 at 710 nm;(2)Normalizing the hyperspectral reflectance data of the black soil partly eliminates the noises introduced when testing different samples;(3)The coefficients of determination R² of the models are all higher than 0.9 whereas their RMSEs are all lower than 2.1,when the maximum and minimum organic matter contents are 117.8 and 6.5 mg kg^-1,so the models are quite good in stability and predictability,and can be used in fast testing of organic matter in black soil;(4)The normalized first derivate model is the optimal one.