Although depending on a wide array of intricate phenomena, tumor growth results, at the macroscopic scale, in relatively simple time curves that can be quantified using mathematical models. Here we assessed the descriptive power of the most classical of these and identified what models are best adapted, based on data from two different experimental settings. We also assessed the predictive power of these models and showed that the most descriptive ones were not necessarily the most predictive and had limited prediction accuracy when no statistical information is used about distribution of the models parameters in the population, for one of the experimental data set, while the other had a linear profile that allowed good predictability. When only few data points were used, we analyzed a method that takes into account the distribution of parameters in a given database for individual estimation of the models parameters. It revealed very helpful and significantly improved the prediction success rates, differentially among the models. These results could be of value for preclinical cancer research by suggesting what model is best adapted when assessing anti-cancer drugs efficacies. They also offer clinical perspective on what can be expected from mathematical modeling in terms of future growth prediction.