Although there is extensive information concerning the colonization sequences of benthic communities, little is known about the successional development of subtidal hard-bottom habitats in highly productive coastal upwelling areas. In these systems, succession is predicted to be fast due to high growth rate of the later dominant colonizers. Using artificial hard substrata a field experiment was conducted in a rocky subtidal area off Northern Chile (Humboldt Current System) and monitored at 3-month intervals to test the following hypotheses: (i) epibenthic succession may proceed through consecutive replacement of species, (ii) there is a fast convergence rate towards natural communities, and (iii) different seasonal starting points for the colonization will produce different community structure over a 1-year period of exposure. Panels were installed on a vertical wall at 17m water depth. Three replicate panels were sampled every 3months over a period of 27months. As a reference, six haphazardly selected plots from the surrounding natural community were surveyed at each sampling date. To evaluate how seasonally varying substratum availability affects community development, further panels were exposed for a 12-month period, starting in four different seasons (n=3 replicates per season). Community succession was slow and occurred through progressive changes, between early encrusting red corallines, middle Balanus flosculus and late Lagenicella variabilis. After 27months, the community composition, but not its structure, was similar between experimental and reference communities on surrounding rocky bottoms. Seasonality had no effects and after 1year of exposure the experimental communities converged towards a common structure. This study indicates that succession of subtidal epibenthic communities follows a slow and predictable pattern with a dominant late colonial species. In addition, aseasonal variability might be more relevant during colonization and succession in this upwelling ecosystem.