Background: The titration of opioid dosage is necessary for adequate pain relief with acceptable side effects among individuals with cancer pain. The titration process can be achieved by non-patient administration or PCA pump. The aim of this study was to evaluate the efficacy of PCA versus non-PAC titration for severe cancer pain. Methods: Patients with severe cancer pain (NRS ≥ 7/10 at rest) were randomized into PCA or non-PCA titration and stratified by opioid tolerance or intolerance. For PCA, the pump was set as no continuous dose, hydromorphone bolus dose was 10%-20% of the total equianalgesic of past 24h for opioid tolerance, or 0.5 mg for opioid intolerance. The lockout time was 15 min. For non-PCA, initial hydromorphone bolus was the same with PCA. Reassess pain at 15 min. The dose of hydromorphone was increased by 50%-100% if pain unchanged or increased, or repeated if NRS was 4-6, or continue at current dose as needed if NRS≤3. The primary endpoint was the time to successful titration (TST) - time from start to the time of pain controlled at NRS ≤ 3 in two consecutive evaluation with 15-min intervals, which was tested by K-M curve. Results: A total of 214 patients were randomized (106 in PCA, 108 in non-PCA) in 17 study sites. The most common sites of primary cancer were lung (21.03%), stomach (15.89%), colorectal (14.49%) etc. Median TSTs were 0.50h in PCA, 0.79h in non-PCA, HR 1.64 (95% CI 1.23, 2.17, P = 0.00127). In opioid tolerance, 0.50h in PCA, 1.00h in non-PCA (HR 1.92, 95% CI 1.32, 2.78, P = 0.0025). while in opioid intolerance, 0.50h in PCA and 0.50 in non-PCA (HR 1.35, 95% CI 0.88, 2.04, P = 0.162). The median dosage of hydromorphone for TST was 1.00mg (P₂₅, P₇₅ 0.50, 2.00) in PCA, 1.50mg (P₂₅, P₇₅ 1.00, 2.50) in non-PCA (P = 0.086). In opioid tolerance, 1.00mg (P₂₅, P₇₅ 1.00, 2.00) in PCA, 2.00mg (P₂₅, P₇₅ 1.00, 4.00) in non-PCA (P = 0.009). In opioid intolerance, 1.00mg (P₂₅, P₇₅ 0.50, 2.00) in PCA and 1.00 mg (P₂₅, P₇₅ 0.50, 2.00) non-PCA (P = 0.793). Mean patient satisfaction assessed by ESAS score was significantly superior in PCA to non-PCA (0.62±0.67 vs 1.27±0.98 for ITT, 0.66±0.66 vs 1.39±1.00 for opioid tolerance, and 0.56±0.69 vs 1.13±0.95 for opioid intolerance). Adverse events were similar in both PCA/non-PCA groups. Conclusions: PCA IV hydromorphone titration provided quicker analgesic effect, higher patients satisfaction, and a similar tolerability as compared to non-PCA administration in patients with severe cancer pain. Clinical trial information: NCT03375515.