Background Access to human hosts by Anopheles mosquitoes is a key determinant of vectorial capacity for malaria, but it can be limited by use of long-lasting insecticidal nets (LLINs). In Malawi, pyrethroid-treated LLINs with and without the synergist piperonyl butoxide (PBO) were distributed to control malaria. This study investigated the blood-feeding patterns of malaria vectors and whether LLINs containing pyrethroid and PBO led to a reduction of human blood feeding than those containing only pyrethroids. Methods Mosquitoes were sampled inside houses from May 2019 through April 2020 by aspiration, pyrethrum spray catch, and light trap methods in two sites. One site (Namanolo, Balaka district) had LLINs containing only pyrethroids whereas the other (Ntaja, Machinga district) had LLINs with both pyrethroids and PBO. Anopheles species, their blood-meal host, and infection with Plasmodium falciparum were determined using PCR methods. Results A total of 6585 female Anopheles were sampled in 203 houses. Of these, 633 (9.6%) were blood-fed mosquitoes comprising of 279 (44.1%) Anopheles arabiensis, 103 (16.3%) Anopheles gambiae 212 (33.5), Anopheles funestus, 2 (0.3%), Anopheles parensis and 37 (5.8%) were unidentified Anopheles spp. Blood meal hosts were successfully identified for 85.5% (n = 541) of the blood-fed mosquitoes, of which 436 (81.0%) were human blood meals, 28 (5.2%) were goats, 11 (2.0%) were dogs, 60 (11.1%) were mixed goat-human blood meals, 5 (0.9%) were dog–human, and 1 was a mixed dog-goat. Human blood index (fraction of blood meals that were humans) was significantly higher in Namanolo (0.96) than Ntaja (0.89). Even though human blood index was high, goats were over-selected than humans after accounting for relative abundance of both hosts. The number of infectious Anopheles bites per person-year was 44 in Namanolo and 22 in Ntaja. Conclusion Although LLINs with PBO PBO may have reduced human blood feeding, access to humans was extremely high despite high LLIN ownership and usage rates in both sites. This finding could explain persistently high rates of malaria infections in Malawi. However, this study had one village for each net type, thus the observed differences may have been a result of other factors present in each village.