Objective To investigate the effectiveness and safety of four dosage regimens of patient controlled analgesia (PCA) on postoperative analgesia in children after pediatric neurosurgery. Methods According to the random number table method, a total of 160 patients, aged 7 to 12 years old, American Society of Anesthesiologists (ASA) Ⅰ‒Ⅲ, scheduled for neurosurgery were di⁃ vided into four groups (n=40): a control group (group C), a fentanyl group (group F), a morphine group (group M) and a tramadol group (group T). All the patients received PCA with different dosage regimens. Group C was given normal saline; group F was infused with a loading dose of 0.5 μg/kg, with a single bolus dose of 0.1‒0.2 μg/kg and a background dose of 0.1‒0.2 μg·kg−1·h−1; group M was infused with a loading dose of 40‒50 μg/kg, with a single bolus dose of 10‒20 μg/kg and a background dose of 1‒4 μg·kg−1·h−1; and group T was given a loading dose of 500 μg/kg, with a single bolus dose of 100‒200 μg/kg and a background dose of 100‒400 μg·kg−1·h−1. The total volume of analgesia pump was adjusted to 100 ml with normal saline and 0.4 mg/kg of ondansetron was added. The bolus locking time was 15 min. Then, postoperative pain scores were evaluated 1, 2, 4, 8, 16, 24, 36 h and 48 h after surgery using the Wong⁃Baker Facial Scale (WBFS) and the Numeric Rating Scale (NRS); and adverse reactions such as change in consciousness and respiratory de⁃pression were recorded. Results Significantly higher WBFS scores and NRS scores were observed in group C 1, 2, 4, 8 h and 16 h after surgery than those in group T, F and M (P<0.05). Meanwhile, significantly lower WBFS scores and NRS scores were observed 1, 2, 4, 8 h and 16 h after surgery in group M than those in group F and T (P<0.05). Group T presented an increased incidence of nausea and vomiting 48 h after surgery, compared groups C, M and F (P<0.05). The average total doses used in PCA pump were lower in groups F and M than that in group T (P<0.05). Group C presented an increased rescue analgesia rate 48 h after surgery than groups F, M and T (P <0.05). There is no significant difference in other indicators among the groups. According to multiple regression analysis, occipital cra⁃ niotomy [odds ratio (OR)=0.295, 95%CI 0.093‒0.937, P=0.038] and the dosage regimen of group M (OR=0.114, 95%CI 0.022‒0.599, P =0.010) were the protective factors that reduce the risks of moderate to severe postoperative pain. Conclusions PCA analgesia with morphine produce the best analgesic effects for children aged 7‒12 y 48 h after neurosurgery, with is safe and effective, without chang⁃ es in the incidence of adverse reactions.