Purslanes are frequently used in Brazil as potted plants and for landscaping because of their beautiful colored flowers. Despite their commercial importance, the genetic diversity within ornamental purslane remains largely unknown. Thus, a complete characterization of ornamental purslane resources is essential for its utilization since genetic variability is indispensable for the efficient development of new cultivars. This study aimed to evaluate genetic variability and recommend accessions to start a genetic breeding program for purslane (Portulaca umbraticola). Twenty purslane accessions were collected in public places in Areia City and Santa Rita City, Paraiba State, Brazil. The experiment was conducted in a greenhouse. An entirely randomized experimental design with 20 accessions and 5 replicates was utilized. Plants were evaluated for quantitative and qualitative traits. Data were submitted to analysis of variance and grouped by Scott–Knott’s criteria (p ≤ 0.05). The accessions were grouped using Tocher’s method based on the generalized Mahalanobis distance. Principal component analysis was also used to analyze the genetic divergence. There were significant differences (p ≤ 0.05) for all evaluated traits except for internode distance, plant height, and leaf length. The Scott–Knott criteria clustered the accessions into two groups for all traits except days to flowering (five groups). Tocher’s grouping based on Mahalanobis distance allowed the accessions to be separated into eight clusters. The genetic diversity observed in this study was supported by the consonance between the Tocher cluster method and the PC analyses. The traits that most contributed to genetic divergence among accessions were days to flowering (70.37%), flower length (6.10%), leaf width (3.89%), branch number (3.57%), and stem length (3.36%). Considering the criteria for use in outdoor spaces, wider plants with several large colored flowers, thicker stems, greener leaves, and earliness were interesting. Accessions 5, 7, 13, 14, 16, 18, and 20 should be selected as parents for further breeding programs. However, if the objective is to produce potted plants, dwarf plants with bigger leaves are desired, and accession 1 should be selected. This will create greater variability in segregating populations, which will increase the overall quality of ornamental use of this species. For those accessions with no reproductive verticils but with other excellent attributes, for example, high chlorophyll content breeding strategies, such as induced mutations or protoplast fusion, can be used.