Objective: To analyze the efficacy and safety of pulsed radiofrequency (PRF) for the treatment of thoracic postherpetic neuralgia (PHN) in elderly patients with different pain phenotypes. Methods: A total of 201 elderly thoracic PHN patients, including 110 males and 91 females aged (72.2±6.9) years who received high-voltage, long-duration PRF at the dorsal root ganglion at Nanjing Drum Tower Hospital Clinical College of Nanjing University of Chinese Medicine from January 2020 to December 2022, were retrospectively included. The neuropathic pain symptom inventory (NPSI) was used to evaluate the five different pain phenotypes, which included superficial spontaneous pain, deep spontaneous pain, paroxysmal pain, evoked pain, and paresthesia/dysesthesia, and to analyze the distribution of the five pain phenotypes. The numerical rating scale (NRS) and NPSI scores of all patients were compared before treatment and three months after treatment to evaluate the efficacy and safety of PRF for different pain phenotypes and pain phenotype combinations. Results: All patients had two or more pain phenotypes, and 50.2% (101/201) of the patients had five pain phenotypes at the same time. Compared with those before treatment, three months after treatment, the NPSI scores for superficial spontaneous pain, deep spontaneous pain, paroxysmal pain, evoked pain and paresthesia/dysesthesia decreased (all P<0.05), and the scores decreased by[M(Q1,Q3)]3.0 (2.0, 4.0), 1.5 (0.5, 2.5), 3.0 (2.5, 4.0), 2.3 (1.0, 4.0), and 1.0 (0.5, 2.0) points, respectively, the differences were statistically significant (P<0.001). The decrease in the NPSI score in patients with paroxysmal pain was greater than that in patients with the other 4 pain phenotypes (all P<0.05). After treatment, the NRS score decreased by 4.0 (3.0, 5.0), 4.0 (3.0, 5.0), 4.0 (3.0, 5.0) and 5.0 (4.0, 6.0) points in patients with 2, 3, 4 and 5 pain phenotypes, respectively, and the difference was statistically significant (P<0.001). The decrease in the NRS score was greater in patients with a combination of 5 pain phenotypes than that in patients with a combination of 3 and 4 pain phenotypes (all P<0.05). No complications, such as pneumothorax, haematoma or infection, occurred in any of the patients during treatment. Conclusion: PRF has different therapeutic effects on PHN patients with different pain phenotypes, it has the best effect on paroxysmal pain, and the treatment is safe.