This paper presents a new methodology to estimate power capacity profiles for smart buildings. The capacity profile can be used within a demand side management system in order to guide building temperature operation. It provides a trade-off between the quality of service perceived by the end user and the requirements from the grid in a demand response context. A data fitting approach and a multiclass classifier are used to compute the required profile to run a set of electric heating and cooling units via an admission control module. Simulation results are reported to validate the performance of the proposed methodology under different conditions, and a comparison made with neural networks in a real world-based scenario.