Determination of biochemical methane potential (BMP) by fermentation tests is time-consuming and costly, and therefore not useful for operators of full-scale biogas digesters. A great advantage of NIRS in determining BMP is the reduction in measurement time from at least one month for chemical determination to a couple of minutes for production of near infrared spectroscopy (NIRS) spectra. An innovative NIRS method that can be used as an alternative to current BMP tests was developed in this study. We tested the Partial Least Squares (PLS) model for a rapid determination of BMP using NIRS by applying a series of pre-processing methods with caution. A total of 88 plant biomass samples of a wide variety were used for model prediction. The standard error of the best PLS model was 37 CH4 NL kg−1 VS, where BMP of the test set was between 136.2 and 477.9 CH4 NL kg−1 VS. Coefficient of determination (R2) and residual prediction deviation (RPD) were 0.84 and 2.49, respectively. This shows that the new NIRS model is moderately successful in application and could be an alternative modern tool to overcome the problems of current BMP methods.