Objective To explore the composition and dynamic changes of symptom clusters of patients with gynecological malignancies at different perioperative time points, to construct symptom networks, to explore core symptoms, and to provide references for the development of precise symptom management plans. Methods Convenient sampling method was used to select 260 patients receiving surgery for gynecological malignancies (cervical cancer, ovarian cancer, endometrial cancer) for the survey. The general information questionnaire and the Chinese version of MD Anderson Symptom Inventory-Perioperative Gynecological Malignant Tumor Module were used to conduct a longitudinal survey of the patients at 3 time points:1 day before the surgery (T1), the 3st day after the surgery (T2), and the 7th day after the surgery (T3, namely before discharge from hospital). Symptoms with an incidence rate of >20% at the 3 time points were subjected to exploratory factor analysis to extract symptom clusters, and the R language was used to construct a symptom network and analyze the centrality indexes. Results Four symptom clusters were extracted for the perioperative period, namely, pain-fatigue-emotion symptom cluster, digestive symptom cluster, energy deficiency symptom cluster and hot flashes-neurological symptom cluster. Among them, the fatigue-pain-emotion symptom cluster existed persistently in T1-T3，and its composition was stable; the hot flashes-neurological symptom cluster existed in T1-T3, the digestive tract symptom cluster existed in T2 and T3, and the energy deficiency symptom cluster existed only in T2, and the symptom composition of the clusters changed dynamically. In symptom networks, sadness and distress were the core symptoms at all 3 time points in the perio-perative period, and the other centrality indicators changed dynamically. In addition to sadness, hot flashes in T1 and nausea in T2 and fatigue in T3 had the greatest tight centrality and mediating centrality, respectively. Conclusion Healthcare professionals should intervene throughout the whole process for stably existing symptom clusters and core symptoms, and combine symptom clusters with network centrality indicators to develop a more precise management plan and improve patients′ perioperative quality of life.