Accurate diagnosis and prognosis judgment of malignant diseases has always been a difficult problem to be solved in clinical practice, and pathological diagnosis is the gold standard. The traditional pathological diagnosis is mainly to diagnose the cell morphology and tissue structure of the sample through naked visual observation, and is a highly subjective, tedious and unrepeatable process, which has the problems of strong subjectivity and insufficient pathology personnel. The continuous integration of artificial intelligence and traditional pathological diagnosis makes pathological diagnosis gradually become intelligent. The artificial intelligence technology represented by deep learning shows great potential and becomes a powerful diagnostic tool for pathologists. AI-assisted diagnosis system is of great significance in the identification, classification and grading , tumor metastasis and prognosis of benign and malignant cells and histopathology. At the same time, there are corresponding problems, challenges and opportunities in the process of clinical application.