Background: Although biochemical markers have been used to monitor training loads (TL), it is unknown if they can be used to predict muscle injuries (MI) in professional football (soccer) players (PFP). Objectives: To evaluate the relationship between the incidence of MI, serum concentration of creatine phosphokinase (CPK) and urea, as well as TL in PFP. Methods: Twenty-three PFP from a Colombian first-division team were enrolled in a retrospective cohort study. CPK, urea, TL and new MI were measured during 19 weeks. CPK and urea serum levels within 4 weeks before a diagnosed MI were compared to those measured preseason. CPK and urea relationship with TL were analyzed using a mixed-effects model. Results: The subjects had an age of 25.3 ± 4.2 years. Nine subjects presented with MI during follow-up, 66.6% of which were localized to hamstrings. Serum CPK and urea profiles were constructed for each player along the season. Injured players had a significant elevation of these markers within 4 weeks before the injury was clinically evident when compared to their own preseason values. Expected individual increases in CPK and urea according to TL during the season were estimated. Conclusions: Since CPK and urea values rose several weeks before the MI became overt, constructing CPK and urea profiles for each player during the whole preseason and season may help identify peaks in their concentration as early markers of MI. A tight biochemical control of training may become a preventive strategy for MI, but the use of published reference values is discouraged.