Recurrence is a characteristic feature of nontypeable Haemophilus influenzae (NTHi) infection, particularly during COPD exacerbations. This makes the management of COPD more challenging. Recurrent infections are caused either through persistence and relapse of earlier infection or by re-infection with a new strain. For a number of bacterial pathogens, whole genome sequencing has been shown to provide a level of resolution that can differentiate between strains from relapsed and new infections. This presented an opportunity to develop in this study a simple workflow for genome-wide comparison of paired isolates of NTHi. In this pilot study, we examined four pairs of longitudinal isolates of NTHi from the sputum of COPD patients sampled at baseline and recurrence. The isolates were sequenced on an Illumina MiSeq platform and low-quality bases and adaptors were removed from fastq reads using TRIMMOMATIC. Based on in silico multilocus sequence typing (MLST) profile, longitudinal isolates from the same patient were found to be non-unique. The paired isolates were further analysed in respect of the number and location of single nucleotide polymorphisms (SNPs). The genome of the baseline isolate was assembled de novo and was annotated automatically using SPADES and PROKKA, respectively. The sequence reads of re-infection isolates were then mapped to the annotated reference baseline genome, and variants were then called using SNIPPY. For each of the longitudinal isolates, the analysis found that there was little genetic distance (<5 SNPs) between members of each pair of longitudinal isolates i.e. 2, 0, 0, and 0 SNPs. Based on this pilot set of paired NTHi isolates, the data to date suggest that relapse of an earlier infection may be more common than re-infection with a new strain in COPD exacerbations. Further analysis of a larger number of longitudinal isolates is required, but should this trend hold, it would have implications for the effectiveness of antimicrobial therapy in the treatment of COPD exacerbations, particularly in relation to NTHi.