A System And Method For Early Prediction Of Esophageal Cancer Subtypes From Histopathology Images The system comprises a data selection unit for receiving Histopathological H&E-stained images of ADC and SCC and selecting images of a select group of patients; a data pre-processing unit for extracting tiles and normalizing color; a central processing unit for training a deep learning and machine learning-based model(108) to predict Esophageal cancer subtypes from histopathology images, wherein the model comprises: a feature extraction processing unit for obtaining a feature vector comprising steps of a)training the network to learn generic image features;b)capturing general visual patterns; c)obtaining feature vectors; d)fine-tuning the pre-trained ResNet 101 model on esophageal cancer subtype classification dataset; e) extracting features from images; and a prediction processing unit for predicting Esophageal cancer subtype and classifying into adenocarcinoma and squamous cell carcinoma using different Machine learning and Feed-forward neural classifiers, wherein tissue-level labels are employed to represent the WSI using a feature matrix for identifying the Esophageal cancer subtype.