Objective To exploring the predictive value of a clinical and radiomics combined model based on transperineal pelvic floor ultrasound (TPFU) for postpartum pelvic organ prolapse (POP). Methods The retrospective analysis was conducted on the medical records, measurement data of the levator?hiatus (LH) pelvic floor ultrasound, and LH pelvic floor ultrasound images of 331 ?primiparous and multiparous women (participants) in Nanfang Hospital Zengcheng Campus of Southern Medical University from January 2024 to August 2024. According to the POP-Q diagnostic criteria, they were divided into POP group (266 cases) and non POP group (65 cases). Construct a clinical model by combining the medical records of 331 participants with LH ultrasound measurements. In addition, manually delineate the region of interest (ROI) in TPFU images, which is the LH region. Then perform radiomics feature extraction and feature screening, and construct an radiomics model based on the selected features. Finally, the selected clinical features and radiomics features will be combined using logistic regression to construct a combined model. Results In clinical models, the best performing model is the RandomForest(RF) model, with AUC of 0.845 and 0.820 for the training and testing sets, respectively. In the radiomics models, the best performing model is also the RF model, with AUC of 0.865 and 0.831 for the training and testing sets, respectively. The AUC of the training and testing sets of the combined model were 0.897 and 0.856, respectively, which were higher than those of the clinical model and the radiomics model.Conclusion The combined model can better predict the occurrence of POP, providing a basis for early intervention and treatment of POP in clinical practice.