One of the key issues facing the environmental manager is the sustainable management of solid waste. Recovery of this waste for recycling often presents the best practical environmental option and companies, industry and Government have set a number of targets for its recycling. A particular challenge is encountered when waste recovery depends on voluntary action. Currently, there is no means of predicting the performance of these voluntary schemes in advance. This paper presents a new mathematical model of waste recovery that can potentially enable these predictions to be made. The model utilizes a process simulation approach to describe the material balance from consumption to discard. Flow partitions within the balance are governed by human decisions. In the model these decisions are represented by probability-distribution functions comprising an explained component and an unexplained random component. The explained component is linked to the demographics of the sample population. The key antecedent factors that affect the human decisions have been delineated from a literature survey of the psychological factors affecting recycling. Their representation in the model is consistent with established theories of recycling behaviour. The new model is able to provide simultaneous time-series predictions of scheme participation rates, total weights of material collected and the composition of this material. The model has been tested on a newspaper kerbside collection scheme and has produced good fits to all measured indicators of recycling performance. The model has additionally predicted the existence and correct scale of observed local variations in recycling behaviours between streets. The potential use of the model as an environmentalmanagement tool is discussed with reference to establishing new collection schemes and to the optimization of the performance of existing schemes.