Due to difficulties in drug penetration in M. tuberculosis, a prodrug approach based on mycobacterial activation appears as a promising strategy to increase the delivery of antitubercular drugs to the target microorganisms. Esters have been successful used by us and others to deliver drugs to mycobacteria, however because very little is known about the metabolic hydrolysis of esters by mycobacteria in connection with prodrug activation, we decided to study the process further. For that we selected a series of 13 benzoates with different chain lengths and ramifications in the alkoxy side as model prodrugs and examined their hydrolysis by a mycobacterial homogenate, comparing the results with those obtained parallelly in human plasma and in total rat liver homogenate. In all biological media, the benzoates with a linear alkyl group showed a parabolic dependence between log(k) and logP (or the number of carbons of the linear alkyl chain) that reached a maximal value for the n-butyl chain. Considering linear correlations for the total number of compounds between log(k) and chosen descriptors, for mycobacterial esterases, pKa of the leaving alcohol (pKa _LG ) seem to be the most important descriptor. Plasma esterases seem to be quite sensitive to the Taft polarity parameter σ* and also to pKa _LG and less sensitive to steric effects. Liver esterases seem to be more sensitive to the Taft steric descriptor E _S ^c . Lipophilicity correlates weakly with log(k) in all the 3 media, however, is more important when one looks for mycobacterial activation selectivity in relation to plasma metabolism or in relation to liver homogenate metabolism. The importance of lipophilicity increases further when biparametric expressions are considered. We showed that it is easy to activate a wide variety of benzoate esters using a mycobacterial homogenate. The data also suggest that with careful design is possible to obtain tuberculostatic prodrug esters sensitive to mycobacterial hydrolases while reasonably resistant to plasma and liver hydrolysis. One important observation is that mycobacterial hydrolysis is less affected by bulky substituents than liver homogenate or plasma hydrolysis. tert-Butyl is probably the substituent in the alkoxy side that seems more adequate to resist simultaneously plasma and liver metabolism, while allowing activation by mycobacterial esterases. Hexyl is also a good option for the medicinal chemist if a linear alkoxy chain is needed.