Quantitative analysis model of acid value of peanut oil was performed using near-infrared spectrometry combined with different interval partial least squares (iPLS). Acid value of peanut oil were determined by acid-base titration while NIR spectra data were recorded; three improved partial least squares (PLS) methods, including interval partial least squares (iPLS), backward interval partial least-squares (BiPLS) and moving window partial least-squares (mwPLS), were used to find the most informative ranges; PLS regression models of acid value were built based on the optimal ranges.The results showed that the model by mw PLS method had the best predictive ability, the RMSECV and RMSEP were 0.24776 and 0.1315, and the calibration and prediction coefficient were 0.9932 and 0.9969. Rapidly and accurately determination of acid value of peanut oil can be achieved by NIR combined with mwPLS.