The invasion of U.S. east coast salt marshes by common reed (Phragmites australis) and the efforts to remove it and restore marshes to their natural vegetation (Spartina spp.) can directly impact mummichog (Fundulus heteroclitus) as this abundant species is critically linked to marsh habitat. We estimated population density, growth, and biomass to determine production of mummichog in treated Phragmites (“Treated,” now dominated by Spartina), untreated Phragmites, and naturally occurring Spartina habitats in Delaware Bay using throw traps for small fish (mean  =  24.1 mm standard length or SL, 95% CI  =  10–38) and tag/recapture for large fish (mean  =  36.5 mm SL, 95% CI  =  18–64). Mean population density of small fish on the marsh surface was significantly higher in the Spartina (20.2 fish m⁻²) and Treated (14.1 fish m⁻²) habitats than in the Phragmites (0 fish m⁻²) habitat. Population density of large fish was similar among all three habitats (mean  =  0.9–1.7 fish m⁻²). Mean absolute growth rates of large fish were significantly higher in the Spartina (0.24 mm d⁻¹) and Treated (0.24 mm d⁻¹) habitats than in the Phragmites (0.13 mm d⁻¹) habitat. Mean monthly mummichog production during June to September varied among habitats with Spartina highest (1.22 g dw m⁻² mo⁻¹), Treated intermediate (0.51 g dw m⁻² mo⁻¹), and Phragmites lowest (0.07 g dw m⁻² mo⁻¹). Small fish were the largest contributor to the production estimates in Spartina and Treated habitats. The Phragmites habitat had little or no standing water at low tide (i.e., optimal habitat for small fish was lacking), and thus, it had the lowest production for mummichog. These results also indicated that Treated marshes were more similar to Spartina than to Phragmites habitat; therefore, it appears that habitat quality and mummichog production can be increased with restoration.