A methodology for the development of self-structuring fuzzy rule-based (FRB) models of HVAC components is proposed in the paper. Advantages of this method include model transparency (the linguistic rules are easily inspected), the possibility to insert expert knowledge into the model generation, and economy in computational effort in generating model output. The novelty of the proposed methodology lies in the specific approach used for automatic rule extraction . It uses rule indices and therefore significantly reduces the computational load. The rules are extracted from the data without using a priori information about their structure. Modeling a range of ducted fans with different geometrical parameters and cooling coils are considered as illustrative examples. A software program implementing the proposed approach has been developed in the framework of Matlab v.5.2. Simulation data as well as published performance figures are used for FRB models learning. Real experimental data is also planned for use in the near future. The FRB models have potential applications in simulation, control and fault detection and diagnosis.