To address the inefficiencies and inaccuracies in analyzing large-scale medical diagnostic datasets, this paper introduces a deep learning-based method for processing auxiliary medical diagnostic data. The proposed approach involves preprocessing the medical diagnostic data through normalization and principal component analysis to extract relevant features. Subsequently, a neural network utilizing a multilayer perceptron is employed to analyze the preprocessed data, facilitating diagnostic classification. It also provides intelligent support for medical professionals. The method was implemented and tested using the Python programming environment. Results indicate that the proposed approach achieves better performance than other comparative methods and demonstrates significant practical application potential.