<p>JPEG standards are widely used in most picture editing software and devices for image acquisition. Because the image format JPEG is the most popular and utmost used, therefore revealing of manipulation in image forensics for JPEG compression has become self-evident. Hence, in image forensics the identification of legitimacy of JPEG images is substantial research area. In order to solve this issue, this work provides a novel JPEG compression detection methodology on the basis of second order statistics. The fundamental inter-block and intra-block relationships of JPEG quantized Discrete Fractional Cosine Transform (DFrCT) coefficients are modelled using second-order Markov Transition Probability Matrices (MTPMs) statistics. The DFrCT method adds a fractional parameter to the suggested method to improve its efficiency. To determine whether or not JPEG compression is present, the second-order (MTPMs) matrices are calculated and treated as discriminative features. Finally, for classification purposes, the SVM classifier is trained using the generated feature. Extensive tests on the datasets such as BOSSBase and UCID demonstrate that the suggested methodology can detect JPEG compression traces even when anti-forensic attacks are present. Moreover, the experimental results illustrate that the presented technique outperforms numerous existing methods.</p>