Abstract Background Rectocele is a frequent finding in women and is usually asymptomatic. However, it is sometimes associated with symptoms of obstructed defecation syndrome (ODS). While most patients with ODS due to rectocele respond well to conservative treatment, some may require surgical treatment. The aim of the study was to determine the predictors of failure of symptom improvement after rectocele repair. Methods The study included adult women with rectocele who underwent surgical treatment by transperineal repair (TPR) or transvaginal repair (TVR). The preoperative and postoperative assessment was done using the Wexner constipation score, anorectal manometry, and defecography. Results A total of 93 female patients with a mean age of 43.7 years were included. Among them, 65.6% of patients underwent TPR and 34.4% underwent TVR; 22 (23.7%) patients reported failure of significant improvement in ODS symptoms after surgery. The independent predictors of failure of improvement were higher preoperative Wexner score (odds ratio, OR: 1.4, 95% confidence interval, CI: 1.09–1.84, p = 0.009), larger residual rectocele after repair (OR: 2.95, 95% CI: 1.43–6.08, p = 0.003), and lower postoperative maximum tolerable volume (OR: 0.949, 95% CI: 0.907–0.992, p = 0.02). The predictive cutoff point for the preoperative Wexner score was 15. Conclusions Patients with a preoperative Wexner score higher than 15 and larger residual rectocele after surgery may experience little improvement in symptoms after rectocele repair. Although TPR was associated with a poorer relief of symptoms than did TVR; it was not an independent predictor of failure.