Empirical antibiotic treatment with a first-generation cephalosporin plus gentamicin for peritoneal dialysis (PD)-associated peritonitis before culture results are available is still wildly used due to concerns regarding the economic burden and antibiotic-resistant bacterial infections. The aim of this study is to define the factors that predict the outcome of empirical antibiotic therapy for PD peritonitis. This is a retrospective study of patients with PD peritonitis over the last 10 years. Patients who had been treated empirically with intermittent intraperitoneal first-generation cephalosporin and gentamicin were enrolled. Eighty-three patients had 192 episodes of PD peritonitis. In total, 159 peritonitis episodes were treated with intraperitoneal antibiotics combined with first-generation cephalosporin and gentamicin empirically. Twenty-five peritonitis episodes had no pathogens identified by dialysate culture. In total, 122 (122/159, 76.7%) PD peritonitis episodes were caused by bacteria, 9 (9/159, 5.7%) were fungal, and 3 (3/159, 1.9%) were Mycobacterium tuberculosis peritonitis. Sixty-four (64/159, 40.3%) peritonitis episodes were successfully cured by empirical intraperitoneal antibiotic therapy. Empirical antibiotic treatment failed in 95 episodes (95/159, 59.7%). The positive rates of dialysate Gram stain in the empirical treatment success and failure groups were 15.6% and 40.0%, respectively (P = 0.001). The odds ratio of empirical treatment failure with a positive bacteria Gram stain was 3.60 (95% CI: 1.65–7.82). The dialysate Gram staining result was a significant predictor of empirical antibiotics treatment outcome for PD-associated peritonitis. Therefore, using an empirical antibiotic regimen for patients with an initial positive bacterial Gram stain of the dialysate should be introduced cautiously.