Background To investigate the relationship between intrapartum maternal fever and the duration and dosage of patient-controlled epidural analgesia (PCEA). Methods This observational study included 159 pregnant women who voluntarily accepted PCEA. During labor, patients with body temperature ≥ 38 °C were classified into the Fever group, (n = 42), and those with body temperature < 38 °C were classified into the No-fever group (n = 117). The outcome measures included the duration of PCEA, number of PCEA, and total PCEA amount. Body temperature and parturient variables, including interpartum fever status and the duration of any fever were monitored. Results The total PCEA duration and total PCEA amount in the Fever group were significantly higher than the corresponding values in the No-fever group (both, p < 0.05). The duration of fever was weakly correlated with the duration of PCEA (R2 = 0.08) and the total PCEA amount (R2 = 0.05) (both, p < 0.05). The total and effective PCEA were higher in the Fever group than in the No-fever group (both, p < 0.05). The total PCEA duration and total PCEA amount were positively correlated with the incidence of fever (both, p < 0.05). The diagnostic cutoff value for fever was 383 min, with a sensitivity of 78.6% and specificity of 57.3%. The mean temperature-time curves showed that parturients who developed fever had a steeper rise in temperature. Conclusions This study showed that there were weak time- and dose-dependent correlations between PCEA and maternal fever during delivery. A total PCEA duration exceeding 6.3 h was associated with an increase in the duration of maternal intrapartum fever.