Intramuscular fat depot is of major interest for consumers, producers, and the industry. To predict intramuscular (i.m.) lipid deposition in cattle of continental breeds, different models were constructed for different muscles in bulls, steers, and cows. Two independent databases (DB1 and DB2) were developed with homogeneous individual data collected in the same slaughterhouse and total lipids, phospholipids, and triglycerides were analyzed in the same lab with the same procedures. Database DB1 was used with the meta-analysis methodology to fit the predictive models of i.m. lipids, phospholipids, and triglycerides with carcass fatness. Database DB2 was used to evaluate the accuracy of the models predicted. Total lipid and triglyceride contents varied linearly with carcass fatness in bulls, steers, and cows, but phospholipids were more independent of carcass fatness, regardless of the type of cattle studied. In bulls, LM had a lower minimal value (intercept in the model) and greater slope than semitendinosus (ST) and triceps brachii (TB) muscles. In cows, LM showed a greater intercept than ST and TB muscles but a similar slope. In steers, lipid content increased similarly in LM, rectus abdominis (RA) muscle, and ST muscle with carcass fatness. Bulls had a lower intercept than steers but showed a similar trend with carcass fatness. According to the external evaluation using DB2, the models obtained to predict total lipids in LM were more accurate than those obtained in the ST muscle in bulls and cows and in the RA muscle in steers. The models proposed for cows should be used only in the range of carcass fatness used to fit the equations, and further data are needed to fully validate them.