Cyanobacteria produce a vast array of natural products, some of which are toxic to human health, while others possess potential pharmaceutical activities. Genome mining enables the identification and characterisation of natural product gene clusters; however, the current number of cyanobacterial genomes remains low compared to other phyla. There has been a recent effort to rectify this issue by increasing the number of sequenced cyanobacterial genomes. This has enabled the identification of biosynthetic gene clusters for structurally diverse metabolites, including non-ribosomal peptides, polyketides, ribosomal peptides, UV-absorbing compounds, alkaloids, terpenes and fatty acids. While some of the identified biosynthetic gene clusters correlate with known metabolites, genome mining also highlights the number and diversity of clusters for which the product is unknown (referred to as orphan gene clusters). A number of bioinformatic tools have recently been developed in order to predict the products of orphan gene clusters; however, in some cases the complexity of the cyanobacterial pathways makes the prediction problematic. This can be overcome by the use of mass spectrometry-guided natural product genome mining, or heterologous expression. Application of these techniques to cyanobacterial natural product gene clusters will be explored.