Finger millet (Eleusine coracana (L.) Gaertn. subsp. coracana) is the most important millet in eastern Africa and perhaps the oldest domesticated cereal grain in Africa. One of the major factors limiting finger millet production is blast disease caused by the fungus Magnaporthe grisea. Crop wild relatives and landraces present a potential source of novel genes. This study investigated the response of cultivated and wild relatives of finger millet to an isolate of blast disease from western Kenya. Previous germplasm collections were purified through two generations of single‐seed descent before screening alongside improved and farmer‐preferred varieties (FPVs) under a screen house across three seasons. Farmer‐preferred varieties were identified through participatory varietal selection (PVS). The plants were inoculated twice during each growth period using hand‐spraying method and data on disease incidence recorded at grain‐filling stage. Genotypic data was generated using diversity arrays technology (DArT) sequencing and data analysis done using Genstat 18.2 and TASSEL 5.2.58. We observed high heritability (81%), indicating that the variation observed was predominantly genetic. Wild accessions were generally more resistant to the disease in comparison to the cultivated accessions. Preliminary genome‐wide association study (GWAS) using general linear model with principal component analysis led to the identification of 19 markers associated with blast disease that will be be developed into assays for genotype quality control and trait introgression. Wild accessions and landraces of finger millet present a good reservoir for novel genes that can be incorporated into crop improvement programs.