Corn leaf diseases pose significant threats to agricultural productivity and food security, necessitating effective and reliable detection methods. Current manual disease detection methods are time-consuming and often prone to errors, highlighting the need for automated approaches that can efficiently and accurately identify diseased crops. This study aims to develop a robust model for detecting corn leaf diseases by leveraging both machine learning and deep learning techniques.The research process involved data collection from Kaggle, followed by pre-processing steps, including resizing images to 224x224 pixels and applying a color constancy technique to normalize lighting variations. Data augmentation was employed to enhance dataset diversity. Feature extraction was conducted using a pre-trained ResNet50 model with additional max pooling layers. For classification, Support Vector Machine (SVM), Random Forest (RF), and Logistic Regression (LR) were utilized. Four distinct approaches were evaluated: (1) combining data augmentation and color constancy, (2) using data augmentation without color constancy, (3) applying color constancy without data augmentation, and (4) without applying data augmentation or color constancy.The evaluation revealed that the combination of data augmentation and color constancy significantly improved model performance, with the SVM model achieving the highest accuracy of 98.44%. These findings highlight the critical role of advanced pre-processing techniques in enhancing the accuracy of leaf disease detection models and underscore the potential of these methods to improve agricultural disease diagnostics.