Spatial synchrony plays an important role in dictating the dynamics of spatial and stage-structured populations. Here we argue that, unlike the Moran effect where spatial synchrony is driven by exogenous factors, spatial correlation in intrinsic/local-scale processes can affect the level of spatial synchrony among distinct sub-populations, and therefore the persistence of the entire population. To explore this mechanism, we modelled the consequences of spatial heterogeneity in aquatic habitat quality, and that of temporal variation in local extinction probability, on the persistence of stage-structured mosquito populations. As a model system, we used two widely distributed mosquito species, Aedes albopictus and Culex pipiens, both key vectors of a range of infectious diseases. Spatial heterogeneity in aquatic habitat quality led to increased population persistence, and this pattern was more pronounced at intermediate dispersal rates, and in the long-dispersing species (C. pipiens). The highest regional persistence was obtained at high dispersal rates. This is probably because dispersal, in our model, did not carry any additional costs. Population persistence of both species was negatively correlated with increased temporal variation in local extinction probability. These differences were stronger in the short-dispersing species (A. albopictus), especially at intermediate dispersal rates. The dispersal of A. albopictus adults in each time step was limited to the nearest habitat patches, weakening the positive effect of spatial heterogeneity in aquatic habitat quality on population persistence. In contrast, C. pipiens adults could disperse into more remote sub-populations, resulting in much higher recolonization rates. Hence, the negative effect of temporal variation in local extinction probability on patch occupancy disappeared at intermediate dispersal rates. We suggest that effectively controlling these two mosquito species requires making few spatially synchronized control efforts (i.e., generating high temporal variation in local extinction probability), rather than many asynchronized local control efforts. Finally, our model can be easily fitted to other organisms characterized by complex life cycles, and it can be also used to examine alternative scenarios, including the effect of spatial configuration of local habitat patches and dispersal kernel shape on population persistence.