Cybersecurity represents a central topic of the Information Era because of the novel threat developed through Internet diffusion. In particular, the social engineering and disinformation risks increase the necessity of novel methodologies able to discern truth from false. For this purpose, this paper aims to describe a Fake News detection method that takes advantage of machine learning techniques for analyzing specific features. In particular, the proposed method elaborates on syntactic features, acquires news topics, and applies sentiment analysis. Moreover, the developed approach exploits Context Awareness to compare themes typical for Fake News in a temporal context with the news topics. Selected features allow the perceptron algorithm application to understand news reliability. The experimental phase of the proposed detection method takes advantage of three datasets for evaluating the ability to discern true news from Fake News. The obtained results are promising.