This chapter highlights how software agents can be conceived to assist users in their tasks and objectives within the smart city. It presents spatio-temporal guidance mechanism based on the contextual planning system with Learning (CPS-L) structure, consisting of a CPS-structure enriched by the learning information. The chapter aims to design a powerful system that assists users in smart-campus activities within a complex university campus to better interact and adapt to users' needs and expectations. The administrative manager of the smart-campus provides a smart application assisting the users through their respective smart devices. The chapter highlights the agent architecture, called higher order agent (HoA) architecture. The HoA model formally represents the behavior of an ambient intelligence (AmI) agent able to reason, communicate and move. The AgLOTOS is an algebraic language specifically defined for original planning process, to be embedded within the agent software architecture.