The purpose of this paper was to investigate the reconstructive changes in foveal microstructures postoperatively and analyse the visual predictors in eyes with surgically closed traumatic macular holes (TMHs). This retrospective study analysed the data collected from 71 eyes of 71 patients, (62 men and nine women) treated with internal limiting membrane (ILM) peeling and intraocular gas tamponade for the repair of TMHs, following blunt eye injury. The mean age was 30.3 years (range 6-72 years). Associated ocular injuries were vitreous haemorrhage, choroidal rupture, macular epiretinal membrane eyes, traumatic cataract and anterior-angle recession. Mean best corrected visual acuity (BCVA) was 0.146 ±0.102, mean minimum linear diameter (MLD) 477.00 ±196.94μm (range 202.50-1,036.00μm), and mean base diameter (BD) 1,151.95 ±429.02 (range 385.50-2,855.00μm) preoperatively. Six eyes underwent combined phacoemulsification +IOL at the same time. Thirteen eyes with peripheral retinal degeneration or holes were treated with endo laser. The mean follow-up time was 8.79 months (range 1-72 months). The primary hole closure rate was 90.14% (64/71). BCVA improved from 0.146 ±0.102 preoperatively to 0.409 ±0.271 at the final visit (p<0.001). The visual acuity improved ≥2 lines in 67.61% (48/71) cases and 40.85% (29/71) achieved 0.5 or better vision. Vitrectomy achieved good anatomic and functional outcomes in eyes with TMH. In 43 cases of initially closed TMHs with spectral domain - optical coherence tomography (SD-OCT) scans preoperatively and at one and six months postoperatively, the number of eyes with intact external limiting membrane (ELM) increased from 18 (41.86%) at one month to 26 (60.47%) at six months (p=0.08), while the number of eyes with a restored ellipsoid zone (EZ) band remained the same (two eyes: 4.65%). The mean length of the EZ defect progressively decreased postoperatively (p<0.001). Poorer preoperative visual acuity (p=0.002), lower mean macular hole (MH) height (p=0.012), and greater preoperative mean length of EZ defect (p<0.001) were associated with worse visual acuity six months postoperatively, but only the preoperative length of the EZ defect was proved to be a predictor, with the cut off value of 1,800μm. Worse visual acuity and greater mean length of the EZ defect at one month were also associated with worse final visual acuity. Limitations of the study include retrospective nature and short follow-up period.