We present an analysis of the reliability of focal mechanisms obtained through moment tensor inversion at the Irpinia Seismic Network, southern Italy. Our analysis is based on the methods proposed by Zahradník and Custódio (2012) and Sokos and Zahradník (2013). We present two different studies: (1) we compute maps of theoretical focal mechanism resolution for the Irpinia region and (2) we study the reliability of the solutions obtained from waveform inversion of five regional earthquakes. Theoretically, we find that when data error is the dominant source of error, focal mechanism resolution is better close to the spots of higher station density rather than at the center of the network. Using real data, we were able to successfully study four of the five regional events, in spite of the large source–station distances (up to ∼280 km) and significant azimuthal gaps (>319°). We used variance reduction, double‐couple percentage, signal‐to‐noise ratio, condition number, and focal mechanism variability index to assess the quality of the solutions. Our quality assessment is validated by comparison with independent focal mechanism solutions.