Multivariate linear regression (MLR) techniques were used to develop empirical models which are able to predict the formation of the main product and byproducts of the adiabatic benzene nitration process, as a function of the main operating conditions. Experiments carried out in a pilot plant enabled us to reproduce the operating conditions of the industrial process, providing experimental data in the intermediate and fast reaction regimes. The nitrobenzene (MNB) formation was modeled according to the film and Danckwerts mechanistic models, and the results were compared with a MLR model, showing that both approaches are suitable for describing this reaction. Nevertheless, the results stress an improved performance of the MLR model when compared to the mechanistic models, despite its structural simplicity. The statistical models developed for the nitrophenols (NPs), namely for the dinitrophenol and trinitrophenol (DNP and TNP, respectively), describe accurately the formation of these byproducts, overcoming the lack of data on kinetic and physical-chemical properties required by the mechanistic approach. The MLR models can be used for process optimization regarding conversion, productivity, and selectivity. By making use of these models, it was possible to estimate the operating conditions (temperature, 81°C; F B /F N ratio, 1.5; residence time, 1.9 min; nitric acid concentration, 2.6%; sulfuric acid concentration, 64%; interfacial area, 46.7 × 10 3 m 2 ‚m -3) that enable the attainment of a 99.99% MNB yield, with a total NP concentration<215 ppm.