Cancer is a major public health and economic issue faced globally today, significantly affecting human health and life. Due to various constraints, exploring the causal relationship between risk factors and cancer is challenging and not exactly accurate. The advent of Mendelian randomization (MR) effectively addresses these issues, providing new avenues for exploring causal relationships. We downloaded literature related to the application of MR in cancer from the Web of Science Core Collection (WoSCC) from 2005 to October 21, 2024, limiting the document type to articles and the language to English, resulting in a total of 2058 articles. We downloaded them in plain text format and extracted information on countries, authors, institutions, keywords, journals, citation counts, and publication dates, utilizing VOSviewer, CiteSpace, and R language for bibliometric analysis. From 2005 to 2024, the number of publications on the application of MR in cancer has shown a growth trend. China was the most productive country (1305); the University of Bristol was the most prolific institution (213); Smith, George Davey published the most articles in this field (59) with a total citation count of 5344; the most prolific journal in this field is Scientific Reports (71). Chronic diseases and cancer, inflammation and cancer, and sex hormones and cancer are three hot topics in the current research on the application of MR in cancer. In the future, optimizing statistical methods, standardizing research processes, collecting data from a broader range of populations, expanding data scale, and integrating other research methods to enhance research quality will be the development trends of MR in cancer research. In summary, this study employed bibliometric methods to comprehensively analyze the literature on the application of MR in cancer over the past 20 years, evaluating the historical development, current applications, research hotspots, and future trends of MR in the field of cancer.