This paper explores the growing danger posed by crypto-ransomware, a type of malicious software that encrypts files and requests payment to unlock them. The study offers a thorough examination of the development of crypto-ransomware, its unique characteristics, and the current methods for its detection. It examines several detection strategies, such as methods based on behavior, signatures, and machine learning. Systems that use signature-based detection for specific patterns or signatures inside the code of the malware to identify and stop attacks. To identify ransomware-related activity, behavior-based detection algorithms concentrate on observing the behavior of processes and applications. For their ability to identify crypto-ransomware, machine learning-based methods, particularly transfer learning, are highlighted in this paper because transfer learning techniques have become a powerful tool for identifying various types of ransomwares, such as locker and crypto-ransomware variants. These methods improve detection efficiency by utilizing both static and dynamic malware properties. In order to find new ransomware strains, hybrid models that combine transfer learning and Machine learning have showed promising results. The review includes several research that used similar methods and produced encouraging outcomes. The review also includes several research that used similar methods and produced encouraging outcomes. The importance of ongoing research and innovation in this field is emphasized in the paper’s conclusion to keep up with attackers’ ever-evolving strategies. The study aims to give an in-depth overview of the status of the research today, point out the positive and negative aspects of different approaches, and suggest areas for future research.