The present invention provides a molecular type prediction system for ovarian cancer, which mainly includes the following steps: step 1, ovarian cancer mRNA gene expression characteristic data extraction module: obtain ovarian cancer gene expression data; step 2, use skleam for all gene expression data The preprocessing.scale method performs standardization processing, according to the formula Z-scroce=(x-μ) / S 2 , process each piece of mRNA expression profile data into data with a mean value of 0 and a variance of 1 that obeys a normal distribution; step 3, select the main characteristic gene data: use principal component analysis (PCA) and Filter feature selection method; step 4, Use the BP neural network to train the model on the genetic data of N features; step 5, use a certain amount of samples to carry out the verification of the bring-back program, and the present invention can realize automatic machine identification and error reporting by means of ovarian cancer pathological slices, and realize fast and high-accuracy Ovarian cancer molecular typing prediction: Utilizing the system of the present invention to predict ovarian cancer molecular typing can better help improve clinical treatment plans.