Myelodysplastic syndromes (MDS) are a collection of clonal diseases with dysfunctional hematopoietic stem cells, characterized by ineffective hematopoiesis, cytopenias, and dysplasia. MDS is increasingly being recognized as a molecularly heterogeneous disease with variability in clinical phenotype, prognosis, and response to treatment. The clonal evolution in response to therapy remains a challenge in the management of MDS. However, resolution of complex clonal architectures is difficult with current bulk sequencing approaches. High-throughput single-cell genomic profiling enables the resolution of tumor heterogeneity, and may improve clinical diagnosis and treatment monitoring by allowing for characterization and early identification of resistant subclonal populations. To enable the characterization of genetic heterogeneity in tumor cell populations, we developed a novel microfluidic approach that barcodes amplified genomic DNA from thousands of individual cells confined to droplets. The barcodes were then used to reassemble the genetic profiles of cells from next-generation sequencing data. We applied this approach to sequential clinical MDS samples, genotyping the most clinically relevant loci across more than 15,000 individual cells. Additionally, to study effects of subclonal mutations on drug sensitivity, ex vivo functional testing was performed on red blood cell-lysed peripheral blood and/or bone marrow aspirate patient samples. Targeted single-cell sequencing was able to recapitulate bulk sequencing data from both peripheral blood and bone marrow aspirate samples. Furthermore, the single-cell nature of our approach enabled definitive determination of mutational co-occurrence within the same cell. For examples in one sample, bulk sequencing identified mutations in JAK2 and NRAS both at 3% variant allele frequency (VAF), where single-cell analysis suggested that these mutations were mutually exclusive, each defining a distinct subclone. Single-cell sequencing allowed for serial monitoring of clonal evolution, with analysis of sequential samples from this same patient showing increase of NRAS clone from 2.5% to 24.7% at time of disease progression after hypomethylating agent therapy. Furthermore, single-cell analysis was able to identify a distinct subclone characterized by a KRAS mutation (0.4% at initiation of therapy and 6.7% at relapse), missed by serial bulk sequencing. Taken together, our results suggest a greater degree of heterogeneity in MDS samples than suggested with bulk sequencing methods alone, and demonstrate utility of single-cell sequencing for sequential monitoring and identification of resistant clones prior to therapy initiation. We show here that this approach is a feasible, scalable, and effective way to identify and track heterogeneous populations of cells in MDS.