In Brazil, Leishmania braziliensis is the main causative agent of cutaneous leishmaniasis (CL), a neglected tropical disease. CL manifests with varying degrees of severity often with high rates of treatment failure. Understanding the parasite factors that contribute to disease presentation and treatment outcome remains a challenge due to difficulties in isolating and culturing parasites from patient lesions. Here we present the development of selective whole genome amplification (SWGA) for Leishmania and demonstrate that this method enables culture-independent analysis of parasite genomes obtained directly from primary patient skin samples. SWGA proves versatile across various Leishmania species found in different host species, which suggests that this method can be applied to both experimental infection models and clinical studies. Analysis of skin biopsies collected from patients in Corte de Pedra, Bahia, Brazil, showed extensive parasite genomic diversity. Integration of SWGA data with published whole genome data from cultured parasite isolates identified region-specific variants in Northeast Brazil, where treatment failure rates are notably high. We also compared parasite genomes from patients that cured and failed treatment and found unique variants associated with treatment failure. While there is no single variant or closely spaced variants specific to treatment outcome, a potential hotspot region linked to treatment failure emerged. SWGA offers a straightforward method to generate Leishmania genomes directly from patient samples, unlocking opportunities to explore the parasite’s population genomics and its association with clinical phenotypes in the host.