Sieving or screening has been the most important unit operation for industrial separation of solid particles. Sieving performance of batch asphalt mixing equipment has an important effect on final asphalt mixture gradation. A Weibull probability distribution model along sieve surface was developed to investigate the effect of material feed rate on the sieving behaviour of material was developed based on the sieving probability analysis of particle swarm. The residence time of particles on the sieve surface is regarded as a random variable, and the sieving process is converted into a life problem. In addition, in order to ensure the accuracy and stability of asphalt mixture gradation, material is divided into two categories, easy-to-sieve particles (ESP) and hard-to-sieve particles (HSP) according to the order of passing through the sieve. Further, a full-scale test in a highway project was conducted to test and verify the influence of material feed rate on sieving performance of vibrating screen. The results are summarized as following: (1) The overlap region between sieving probability density curves of HSP and ESP indicates a poor sieving result including the mixing bin and channeling bin of hot materials. (2) There is an optimum matching between feed rate and sieve surface length for the purpose of stable sieving performance: If the length of sieve surface is the corresponding abscissa value of the intersection, increasing the feed rate would aggravate channeling bin. Conversely, decreasing the feed rate would aggravate mixing bin. (3) It was also indicated that material feed rate will sensitively affect the mixing and channeling rate of materials and it is better to select the curve intersection or its vicinity on the sieving probability density curves of HSP and ESP as feed rate based on the sieve surface length, and avoid substantial changes in feed rate. (C) 2018 Elsevier B.V. All rights reserved.