Abstract: To address the issues of strong subjectivity and low efficiency of the traditional manual evaluation for oil contents in flue-cured tobacco, a nondestructive test method for oil content based on visible-near infrared-hyperspectral technology was proposed. 634 flue-cured tobacco leaf samples from 22 tobacco-growing provinces（autonomous region）in China were studied to obtain spectral information by a hyperspectral imaging system (400-1 700 nm). Five single preprocessing methods and their integrated preprocessing methods were used to enhance the effective signal, and combined with the adoption of the successful projections algorithm (SPA), characteristic bands were screened and single quantitative prediction models of PLSR and SVR and their fusion models for oil content in flue-cured tobacco were constructed. The results showed that: 1) After the single preprocessing such as MSC, SNV, D1 and their combined preprocessing of the original spectrum, the correlation between the spectral reflectance and the oil content score was significantly enhanced, and the model performance was generally improved. 2) The performance of PLSR models based on the full bands of visible light, near-infrared and visible-near-infrared spectra were generally superior to that of SVR models; 3) The coefficient of determination (R2) for the fusion model of PLSR and SVR based on the visible-near-infrared full band of the validation set was 0.721; 4) The PLSR regression model constructed through screening 95 characteristic bands by SPA after the original spectrum was preprocessed by MA had the highest accuracy, while the SVR regression model constructed through screening 56 characteristic bands by SPA after D1+SS preprocessing had the highest accuracy; 5) The performance of the fusion model based on characteristic band screening by spa was enhanced. The R2 of the validation set was 0.724, and the root mean square error (RMSE) was 0.323. Therefore, nondestructive detection of oil content in flue-cured tobacco can be achieved by the proposed visible-near-infrared spectroscopy technology.