Species response to land-use changes are usually assessed by investigating factors affecting distribution, with a single snapshot in time. However, several processes can lead to a same pattern. Focusing on observed, short-term patterns limits our ability to make inferences about ecological processes and responses to environmental change over time. In this study, we assessed changes in occupancy of two primate species in southeastern Brazil, following a major habitat loss due to implementation of a hydroelectric dam. Occupancy was assessed before dam construction and 11 years after, while explicitly accounting for imperfect detection. We assessed the effect of forest patch size and isolation on occupancy and rates of extinction and colonization, driven by landscape modification. Then we calculated occupancy under metapopulation equilibrium and expected time-lags resulting from non-equilibrium. We compared two primate species inhabiting forest patches, the black penciled marmoset Callithrix penicilatta and the black-fronted titi monkey Calliceb us nigrifrons, with markedly different ecological characteristics. Those differences may explain why occupancy dynamics were driven by distinct elements. A fast response to habitat changes was observed only for marmoset, an opportunistic species. However, non-equilibrium states and the possibility of time-lag effects were observed for titi monkey, a species dependent on forest habitat. Our analyses support the need to establish long term monitoring and assess system vital rates overtime. A single snapshot in time may lead to erroneous interpretations of a species response to habitat alteration.