In vivo use of biodegradable magnesium (Mg) metal can be plagued by too rapid a degradation rate that removes metal support before physiological function is repaired. To advance the use of Mg biomedical implants, the degradation rate may need to be adjusted. We previously demonstrated that pure Mg filaments used in a nerve repair scaffold were compatible with regenerating peripheral nerve tissues, reduced inflammation, and improved axonal numbers across a short—but not long—gap in sciatic nerves in rats. To determine if the repair of longer gaps would be improved by a slower Mg degradation rate, we tested, in vitro and in vivo, the effects of Mg filament polishing followed by anodization using plasma electrolytic oxidation (PEO) with non-toxic electrolytes. Polishing removed oxidation products from the surface of as-received (unpolished) filaments, exposed more Mg on the surface, produced a smoother surface, slowed in vitro Mg degradation over four weeks after immersion in a physiological solution, and improved attachment of cultured epithelial cells. In vivo, treated Mg filaments were used to repair longer (15 mm) injury gaps in adult rat sciatic nerves after placement inside hollow poly (caprolactone) nerve conduits. The addition of single Mg or control titanium filaments was compared to empty conduits (negative control) and isografts (nerves from donor rats, positive control). After six weeks in vivo, live animal imaging with micro computed tomography (micro-CT) showed that Mg metal degradation rates were slowed by polishing vs. as-received Mg, but not by anodization, which introduced greater variability. After 14 weeks in vivo, functional return was seen only with isograft controls. However, within Mg filament groups, the amount of axonal growth across the injury site was improved with slower Mg degradation rates. Thus, anodization slowed degradation in vitro but not in vivo, and degradation rates do affect nerve regeneration.