The incremented popularity of Internet of Things (IoT), thanks to improvements both in hardware and software of sensors over the last years, enables the possibility to monitor and gather any kind of data. Additionally, the arrangement of heterogeneous sensors, capable of perceiving information about their surroundings, into a rich Wireless Sensor Network (WSN), allows the appearance of complex systems in which resources are managed more efficiently. Smart cities, buildings, parkings, emergency services are appearing, where control over energy consumption and better sustainability are coupled with an improvement of the comfort of occupants. In this paper, we address the problem of energy optimization in smart buildings, considering both the planning and operational aspects. Specifically, the first aim is to propose an optimal deployment of the WSN inside a building. For this, we present a model able to identify the optimal locations for different types of sensors and gateways, by optimizing energy consumption while fulfilling connectivity, resource, protection, and clustering coverage constraints. Once the IoT system is deployed, we address the problem of how the building actually functions, according to the behaviour of the occupants. In particular, we propose a Building Management System (BMS) capable of efficiently and automatically manage the building elements using human behavioural models, thus lowering the overall building energy consumption whilst maintaining acceptable levels of comfort.