Antimalarial drug resistance is a substantial impediment to malaria control. The spread of resistance has been described using genetic markers which are important epidemiological tools. We carried out a temporal analysis of changes in allele frequencies of 12 drug resistance markers over two decades of changing antimalarial drug policy in Kenya. We did not detect any of the validated kelch 13 (k13) artemisinin resistance markers, nonetheless, a single k13 allele, K189T, was maintained at a stable high frequency (>10%) over time. There was a distinct shift from chloroquine resistant transporter (crt)-76, multi-drug resistant gene 1 (mdr1)-86 and mdr1-1246 chloroquine (CQ) resistance alleles to a 99% prevalence of CQ sensitive alleles in the population, following the withdrawal of CQ from routine use. In contrast, the dihydropteroate synthetase (dhps) double mutant (437G and 540E) associated with sulfadoxine-pyrimethamine (SP) resistance was maintained at a high frequency (>75%), after a change from SP to artemisinin combination therapies (ACTs). The novel cysteine desulfurase (nfs) K65 allele, implicated in resistance to lumefantrine in a West African study, showed a gradual significant decline in allele frequency pre- and post-ACT introduction (from 38% to 20%), suggesting evidence of directional selection in Kenya, potentially not due to lumefantrine. The high frequency of CQ-sensitive parasites circulating in the population suggests that the re-introduction of CQ in combination therapy for the treatment of malaria can be considered in the future. However, the risk of a re-emergence of CQ resistant parasites circulating below detectable levels or being reintroduced from other regions remains.