Since the 90's, automatic systems take a growing place into our everyday lives. It might be embedded systems, such as in robotic, automotive or avionic systems, telecommunication or transportation systems or energy services, etc. The presence of such systems offers new possibilities, but the price to pay is the increasing risks of software failures which can have dramatic consequences in terms of human lives or prohibitive costs. Manual validation is expensive, may be impossible for large systems, and is permeable to mistakes. The development of automatic tools serving to analyze or to ensure security/safety has thus become crucial to discover and avoid breaches and mistakes in the development of embedded systems. This habilitation document reports on my contributions to formal methods for discrete event systems. More specifically, the following problems are tackled: the diagnosis of discrete event systems for faulty behaviors that can be represented finite automata (these faults can be permanent or transient). A second contribution concerns the automatic control of concurrent systems composed of multiple sub-systems communicating either synchronously or asynchronously by means of fifo channel. In the field of computer security, the third contribution deals with the enforcement and the detection of confidentiality properties (more specifically the notion of opacity) using diagnosis techniques to detect opacity violations or supervisory control theory to restrict the behavior of a system in order to avoid information leakage.