This paper presents an approach based on Machine Learning (ML) to estimate the performance of small antennas in practical applications. As a proof of concept, the case of printed antennas integrated into compact terminals, such as for Internet of Things (IoT) or 5G applications, is considered.