The invention relates to establishment and improvement of a preoperative evaluation and prediction model for satisfactory tumor cell deduction of advanced ovarian cancer. Herein, improved modeling is carried out on the basis of the Sduidan score table; and serological HE4 and ROMA indexes are added, the HE4 calculated according to the data is larger than 264.7 pmol/L, the CA125 calculated according to the data is larger than 545.6 U/ml, the ROMA indexes calculated according to the data are larger than 90.0%, and the prediction scores are given to be one score respectively; and then two different radiology department high-age doctors perform secondary radiology reading scoring to calculate predicted scores of all cases; and finally, an ROC curve is drawn according to a prediction score and an operation result to obtain an area AUC under a subject working curve, and evaluation of the model is carried out to predict the clinical value of the unsatisfactory tumor cell deduction operation of the advanced ovarian cancer patient. The prediction value of the prediction model is higher than that of a Suidan prediction model. When the model is used for predicting the dissatisfactory tumor cell deficit operation, the sensitivity is obviously higher than that of a Suidan score table, which provides a basis with a certain prediction value for clinicians to judge whether the dissatisfactory tumor cell deficit operation can be performed or not.