We tested the effectiveness of predation by the cyclopoid copepod Mesocyclops longisetus (Thiébaud, 1912) in Culicidae and Chironomidae larvae, aiming to test if (i) the introduction of copepods effectively controls mosquito larvae and (ii) the density of copepods is important for ensuring control. We conducted two semi-field experiments: the first involved 14 experimental runs over 75 consecutive days, compared in a randomized block design, four repetitions per treatment in each experimental run (block); and the second experiment involved a total of five experimental runs, lasting 25 and 33 days each. In the first experiment, culicid larvae were preyed on by copepods, especially at high copepod densities (15 copepods per litre), which reduced culicid densities by approximately 61% and 64% for different containers, considering that the environmental temperature declined linearly by about 10ºC from February (summer) to May (autumn) (32ºC to 22ºC). Even the lowest copepod density (5 per liter) produced a substantial reduction in the abundance of culicid larvae when chironomid larvae were abundant. On the other hand, there was no effective control of chironomid larvae. In the second experiment, increases in Culicidae and Chironomidae were concomitant with decreases in copepods. Thus, when the copepods were in high densities, there was no increase in insects, although the larvae of culicids were not eliminated from the experimental units. We conclude that the introduction of copepods in small containers and ponds can be useful for the control of culicid larvae without harming chironomid populations, and densities up to 15 per litre are recommended, although even low densities cause substantial reduction.