Effective monitoring of populations and communities is a prerequisite for ecosystembased management of marine areas. However, monitoring programs often neglect important sources of error and thus can lead to biased estimates, spurious conclusions and false management actions. One such source of error is 'imperfect detectability', i.e. the inability of investigators to detect all individuals or all species in a surveyed area. Although there has been great effort to develop monitoring methods that account for imperfect detectability, the application of such methods in the marine environment is not as apparent as in other systems. Plot sampling is by far the most commonly applied method for biological monitoring in the marine environment, yet it largely ignores detectability issues. However, distance sampling, mark-recapture methods, repeated presence-absence surveys for occupancy estimation, and removal methods do estimate detection probabilities and provide unbiased estimates of state variables. We review these methods and the relevant tools for their application in studies on marine populations and communities, with the aim of assisting marine biologists and managers to understand the limitations and pitfalls associated with some approaches and to select the best available methods for their monitoring needs.