Abstract Biomass is an important renewable energy source, in particular if obtained by residues it becomes even more sustainable. In Italy, residual biomass coming from olive oil industry, i.e. olive pomace, is produced in a significant amount and is concentrated in olive oil extraction sites, making interesting a possible valorisation of these residues. The different extraction processes employed influence the pomace quality and, consequently, it is fundamental to find a rapid technique to assess its physical and chemical characteristics for a correct valorisation. The aim of the work was to develop a NIR-based methodology to obtain in a fast and cheap way information about olive pomace. Several samples (n = 104) were collected in Marche region and analysed according to standards methods. NIR spectra were acquired using both fiberoptic probe and integrating sphere and subsequently were elaborated with multivariate techniques, i.e. principal component analysis (PCA) and partial least square regression (PLS). Results show that information on extraction process and composition of the pomace can be obtained. Prediction models with performance suitable for quality control applications were obtained for moisture and ash contents, whereas gross calorific value model was suitable only for screening application.