Screening is the most important operation for the separation of solid particles. The distribution of particles on the screen surface is an important factor affecting the screening performance. In this paper, a biological neural network (BNN) approach is proposed for modeling the distribution of particles on a vibrating screen surface. The dynamics of each neuron is characterized by the shunting equation, and the neural connection weights are properly defined according to the movement of particles under different vibration and structural parameters. Neural activities can propagate from the particles input neurons to the whole network through adjacent neural connections. When the iterative calculation is stable, the generated neural activity landscape is used to describe the particles distribution state. Discrete element method (DEM) simulations are carried out to obtain the particles screening processes and the corresponding distribution states. Then, the particles distribution models established using BNN are compared with that obtained by DEM simulations, and the similarities between them are improved by optimizing the BNN model coefficients. Similarity analysis results under different screening conditions show that the general correlation coefficient is higher than 0.9, which verifies the feasibility of the proposed BNN approach. Compared with the traditional kinetic and probability models, the BNN approach has obvious advantages in solving the modeling problem when the particles are fed to screen with multiple areas and non-uniform rate. (C) 2021 Elsevier B.V. All rights reserved.