The development of marine renewables has raised concerns regarding impacts on wildlife, and environmental monitoring is often required. We examined 3 mo of continuous passive acoustic monitoring (PAM) data collected at the Tidal Energy Ltd. DeltaStream turbine deployment in Ramsey Sound, UK. We aimed to assess the performance of the PAM system at an operational turbine, describe the 3D movements and behaviours of small cetaceans in the vicinity of the turbine, and model changes in detection rates against temporal and environmental variables. The PAM system was designed to acoustically detect, classify and track porpoises and dolphins via their vocalisations within a ~100 m radius of the turbine. In total, 247 small cetacean encounters were identified from click detections, which were also used to reconstruct the spatial movements of porpoises and dolphins, including close approaches to the turbine. Not all hydrophones were functional, which limited the ability to localise porpoise clicks; the probability of detecting and localising a click decreased by 50% at a range of ~20 m. Mechanical sounds on the turbine may have alerted cetaceans of its presence. In models examining acoustic detection patterns, the tidal state, time of day, low low-frequency noise levels and moon phase best explained the acoustic presence of porpoises. The limited duration of turbine operation yielded insufficient data to understand the effect of turbine rotation on animal presence and movement near the turbine. This is the first description of how small cetaceans behave and move around a tidal turbine, and we present recommendations regarding how PAM can be used to improve environmental monitoring at future tidal energy sites.