This study assessed the effectiveness of using video cameras mounted on commercial trawls for a deep-sea lobster, the New Zealand scampi Metanephrops challengeri, for acquiring information on habitat associations of demersal species and fishing efficiency. Randomly selected frames from 1 h of recording replicated over 12 trawls and conducted over three fisheries management areas (FMA), were analysed for counts of visible scampi, scampi burrows, visible topographical features and visible epifauna and fish taxa. Poisson regression analysis indicated a positive relationship between the distribution of scampi burrows (a proxy of scampi abundance) and conical mounds (formed by bioturbators), but a negative relationship with smooth topography. Counts of scampi burrows were also positively associated with counts of visible scampi for four trawls, while significant relationships were seldom found between counts of scampi burrows and other commonly observed individual taxa. Scampi trawl efficiency was low, with only 56 of the 186 observed scampi (30%) passing over the footrope and into the trawl net. The results confirm the effectiveness of video sampling from commercial trawls to improve our understanding of habitats and fishing efficiency in deep-sea fisheries.