The aim of this review is to summarize the current scientific knowledge about acceleration and deceleration demands during football training. A systematic search of three electronic databases (PubMed, SPORTDiscus, Web of Science) was performed to identify peer-reviewed relevant English-language articles, following PRISMA guidelines. All acceleration and deceleration data were analyzed and organized into four categories: i) training drills variables (i.e. manipulated drills variables such as number of players in small-sided games), ii) training exercises (i.e. different drills such small games or circuit training), iii) players’ positions (i.e. demands for each playing position) and iv) training schedule (i.e. training sessions presented as microcycles, season sections or full season). Full-text articles of 42 studies were included in the final analysis. Players’ level included: amateur, youth, semi-professional, professional and elite players. All playing positions were considered, including goalkeepers. Six different global position systems brands were used, with the majority measuring data at 10 Hz. Different thresholds and intensities were used in several papers. Lower acceleration and deceleration intensities occurred more often than higher intensities in all four categories. Different exercises elicit different demands and small-sided games presented higher acceleration and deceleration demands than circuit training and other running based drills. Furthermore, manipulating drills variables, as reducing or increasing number of players in small-sided games increase or decrease demands, respectively. Additionally, wide playing positions, such as fullbacks, are generally exposed to higher acceleration and deceleration demands. From a planning point of view, acceleration and deceleration demands decrease as match day approaches.