A new generation of strategies is evolving that aim to block malaria transmission by employing genetically modified vectors or mosquito pathogens or symbionts that express anti-parasite molecules. Whilst transgenic technologies have advanced rapidly, there is still a paucity of effector molecules with potent anti-malaria activity whose expression does not cause detrimental effects on mosquito fitness. Our objective was to examine a wide range of antimicrobial peptides (AMPs) for their toxic effects on Plasmodium and anopheline mosquitoes. Specifically targeting early sporogonic stages, we initially screened AMPs for toxicity against a mosquito cell line and P. berghei ookinetes. Promising candidate AMPs were fed to mosquitoes to monitor adverse fitness effects, and their efficacy in blocking rodent malaria infection in Anopheles stephensi was assessed. This was followed by tests to determine their activity against P. falciparum in An. gambiae, initially using laboratory cultures to infect mosquitoes, then culminating in preliminary assays in the field using gametocytes and mosquitoes collected from the same area in Mali, West Africa. From a range of 33 molecules, six AMPs able to block Plasmodium development were identified: Anoplin, Duramycin, Mastoparan X, Melittin, TP10 and Vida3. With the exception of Anoplin and Mastoparan X, these AMPs were also toxic to an An. gambiae cell line at a concentration of 25 µM. However, when tested in mosquito blood feeds, they did not reduce mosquito longevity or egg production at concentrations of 50 µM. Peptides effective against cultured ookinetes were less effective when tested in vivo and differences in efficacy against P. berghei and P. falciparum were seen. From the range of molecules tested, the majority of effective AMPs were derived from bee/wasp venoms.