I studied habitat use by semi-aquatic snakes on a reclaimed strip mine in northeast Texas. I trapped snakes of the genera Agkistrodon and Nerodia at ten study ponds during two years. Additionally, five adult male Nerodia erythrogaster and five adult male N. rhombifer were tracked using radiotelemetry to compare habitat use between these two species, the most abundant snake species at the study site. The intraspecific density of N. erythrogaster and N. fasciata did not differ statistically among the ponds. The density of N. rhombifer was not uniform among the ponds, but mining history, pond age, shoreline length, and shoreline vegetation did not explain the observed among-pond differences. However, when the data from all ponds were combined, more N. rhombifer were captured in vegetated than unvegetated areas. The density of Agkistrodon piscivorus was not uniform among the ponds; the presence of a forested bottomland abutting one pond may have accounted for its high density there. Radio-tagged N. rhombifer were more closely associated with water than were radio-tagged N. erythrogaster. In the summer, N. rhombifer spent more time near permanent ponds than did N. erythrogaster, which spent relatively more time near temporary bodies of water. In the autumn, however, as N. rhombifer increased its use of temporary water sources, the use of permanent ponds vs. temporary sources did not differ significantly between the species. The results of this study indicate that semi-aquatic snake populations can recover quickly after surface mining, provided that ponds and other low-lying areas are created in the reclamation process and that there is a nearby source pool for colonization.