<p>With the rapid development of modern digital imaging techniques, it is possible to obtain big data to evaluate various properties of rocks across disparate scales. Various machine learning (ML) and deep learning (DL) algorithms have emerged as one of the most effective ways to analyze big data from rock digital images. The rock engineering field has witnessed the rapid developments of ML and DL algorithms, which provide large analysis data of rocks to enable researchers to solve various rock engineering problems. Meanwhile, the existing knowledge of rock problems can also develop and enlarge novel ML and DL methods. Thus, it is imperative to select and employ advanced ML and DL algorithms to different rock engineering problems. The purposes of this review are to clearly define the multiscale analysis of rocks, to provide a comprehensive review of the concept and model evaluations with different metric indexes of ML and DL algorithms, and to elucidate their applications on the petro-physical parameter evaluations, realistic 3D modeling, and cracking behaviors as well as failure mechanism analyses of multiscale rocks. In addition, the challenges and future directions of the rapidly developing ML and DL algorithms are also discussed.</p>