Abstract <p>This study aimed to develop a predictive model of crop rotation productivity using adaptive neuro-fuzzy inference. The research utilized data from long-term field experiments conducted by the Siberian Federal Centre of Agricultural Biotechnology of the Russian Academy of Sciences in the Novosibirsk oblast between 1999 and 2019, covering nine types of crop rotations designed for grain production. The methodology incorporated an artificial neural network training algorithm that employed a hybrid optimization method, which combined the least squares method and the backpropagation error method. This approach facilitated the formulation of fuzzy rules with appropriate membership functions based on both input and output data. The adaptive neuro-fuzzy inference system (ANFIS) model for crop rotation productivity was developed using the MATLAB environment. The rules generated during the ANFIS training process accurately determined significant factor combinations that influence the productivity of the defined crop rotations. Forecast modeling for three types of crop rotations revealed the considerable impact of winter crops and crop sequence elements on rotation stability under unfavorable atmospheric moisture conditions, as well as on the efficiency of agrochemical application. A comprehensive analysis based on various accuracy metrics (coefficient of determination 0.78; root mean square error 5.66; mean absolute error 4.31; mean absolute percentage error 20.07%) confirmed the model’s strong predictive capability. The developed ANFIS model demonstrates a high capacity to account for complex, nonlinear relationships among variables affecting crop rotation productivity, and it may serve as a valuable tool for making informed production decisions in both short-term and long-term planning scenarios.</p>