The importance of habitat structural complexity in the assessment and understanding of Marine Protected Area (MPA) performance remains unresolved. Habitat structural complexity is known be an important driver of the abundance and diversity of fishes so therefore may confound spatial comparisons between reserves and non-reserves. This is specially the case when assessments of MPA effectiveness involve non-random allocation of treatments or low replication. Here we present an assessment of the performance of marine reserves in the Lord Howe Marine Park (LHIMP), NSW, Australia while accounting for underlying variation in habitat structural complexity. The LHIMP comprises of two marine reserves where fishing is prohibited, surrounded by habitat protection zones that allow recreational fishing and charter boat fishing operations. High resolution multi-beam surveys in the LHIMP have revealed substantial variation in habitat structural complexity across all zones. As a result, the LHIMP provides an ideal opportunity to investigate the importance of how habitat structural complexity influences the assessment of MPAs. In this study we tested the performance of LHIMP by comparing the abundance of Seriola lalandi (Yellowtail Kingfish) and Carcharhinus galapagensis (Galapagos Whalers) between reserves and non-reserves while accounting for habitat structural complexity. We found that Seriola lalandi abundance was significantly greater in reserves compared to non-reserves but only at sites with high structural complexity. In contrast, the abundance of Carcharhinus galapagensis did not differ between reserves and non-reserves nor with habitat structural complexity. We propose that the disparity in results is due to differences in fishing pressure and habitat preferences between the two taxa. Carcharhinus galapagensis is a pelagic species displaying little affinity to seafloor habitats and is not intentionally targeted by fishers. In comparison, Seriola lalandi displays a stronger tie to benthic habitats and is heavily exploited by fishers. Our findings demonstrate that habitat structural complexity is important to consider when testing MPA performance. We suggest that future studies assessing MPA performance to incorporate measures of habitat variability if the data is available.