This paper discusses the Data-Based Mechanistic (DBM) approach to modelling the micro-climate in agricultural buildings. Here, the imperfect mixing processes that dominate the system behaviour during forced ventilation are first modelled objectively, in purely data-based terms, by continuous-time transfer function relationships. In their equivalent differential equation form, however, these models can be interpreted in terms of the Active Mixing Volume (AMV) concept, developed previously at Lancaster in connection with pollution transport in rivers and soils and, latterly, in modelling the micro-climate of horticultural glasshouses. This can be compared with the incomplete mixing and control volume concepts that have been investigated previously at Leuven. The data used in the initial stages of the research project, as described in the paper, have been obtained from a series of planned ventilation experiments carried out in a large instrumented chamber at Leuven. The overall objectives of this collaborative study are twofold: first, to gain a better understanding of the heat transfer and micro-climate dynamics in the chamber; and second, to develop models that can form the basis for the design of optimal Proportional-Integral-Plus (PIP-LQ) climate control systems for livestock buildings of a kind used previously for controlling the micro-climate in horticultural glasshouses. Although not specifically directed at glasshouse systems, the techniques described in the paper can be applied straightforwardly within a glasshouse context.