Half of encephalitides, which are among the most severe conditions in infectious diseases, remain of unknown etiology. Non-detected infectious etiology in the most severe cases leads to misusage of immunosuppressive therapies. Our experience showed that around 30% of the undiagnosed cases of human encephalitis can be due to inefficient identification of known, or most often unexpected, or new pathogens. In this chapter, we present the methods, from histopathological criteria to molecular biology and bioinformatics, that we routinely use in our clinical metagenomics practice, leading to the identification of new or unexpected pathogens responsible for human encephalitis.