The tsunamigenic potential of an earthquake depends on its size, source depth and focal mechanism. The Hellenic Subduction Zone (HSZ) has been selected in the paper to study this important issue. The HSZ was ruptured by 11 strong (Mw6.0) earthquakes in the time period 2009–2023. One earthquake ruptured onshore but only three out of ten offshore earthquakes produced tsunamis: 1 July 2009 (Mw6.4), 25 October 2018 (Mw6.8), 5 May 2020 (Mw6.6). For each one of the two more recent earthquakes of 5 May 2020 (tsunamigenic, thrust faulting) and 12 October 2021 (non-tsunamigenic, strike-slip faulting) we developed heterogeneous fault models from the inversion of teleseismic P-waveforms, and homogeneous fault models from published moment-tensor solutions. For each fault model tsunami generation and propagation was numerically simulated based on an advanced phase-resolving wave model with the use of higher-order Boussinesq-type equations. The modelled tsunami mareograms are consistent with tide records of the small tsunami (height ∼30 cm) produced by the 2020 earthquake. For the 2021 earthquake the modelled mareograms showed tsunami-like disturbance with amplitude not exceeding the noise level. The tsunamigenic earthquakes of 2009, 2018 and 2020 shared magnitude Mw≥ 6.4, shallow depth (h<20 km), moderate-to-high dip-angle and thrust faulting or oblique slip with significant thrust component. In the remaining seven non-tsunamigenic earthquakes, including the 2021 one, at least one of these features is missing. The results obtained help to better understand the seismic mechanisms of tsunami generation in the HSZ. Further investigation is needed for the central HSZ segment to the south of Crete Island, which historically has not been ruptured by large (Mw>7.0) tsunamigenic earthquakes. In contrast, the western and eastern HSZ segments ruptured by the large 365 AD and 1303 AD tsunamigenic earthquakes.