Digital image forgery has emerged as a significant challenge in multimedia forensics due to the rapid advancementofimage editing software and artificial intelligence-based manipulation tools. In this work, a lightweight image forgerydetectionframework based on hybrid feature extraction and machine learning classification is proposed for distinguishingbetweenauthentic and forged images. The developed framework integrates Discrete Wavelet Transform (DWT)-based frequencyanalysis,Gray Level Co-occurrence Matrix (GLCM) texture descriptors, statistical image features, and edge density analysis tocapturehidden inconsistencies introduced during image tampering operations. Initially, the input images obtained fromtheCASIAv2benchmark dataset undergo preprocessing operations including resizing and grayscale conversion. Subsequently, DWTdecomposition is performed to extract frequency-domain characteristics, followed by GLCM-based texture analysis andstatisticalfeature extraction. Edge-based structural analysis is additionally incorporated using the Canny edge detectionalgorithmtoimprove forgery discrimination capability. The extracted features are fused into a unified feature vector and classifiedusingaRandom Forest classifier. Experimental analysis performed in MATLAB R2024b demonstrates that the proposed framework achieves high classification performance with an overall accuracy of 98% while maintaining low computational complexity suitable for CPU-based implementation. The confusion matrix and feature analysis further confirm the effectiveness of the proposed hybrid feature extraction strategy for digital image forgery detection. Keywords: Image Forgery Detection, Digital Image Tampering, DWT, GLCM, Random Forest, CASIADataset, MachineLearning