The field of Alzheimer’s disease (AD) research has been quite fortunate – in contrast to some other neurodegenerative psychiatric diseases – in that a number of animal models have been developed based on genetic and neuropathological information. These animal models have been “validated” based on the fact that they reflect one or a few of the neuropathological features found in postmortem brain of AD patients. However, the predictive value of these models for drug discovery has been far from spectacular. Limitations of these models include the failure to capture the dynamics of the ongoing pathology in a clinical setting and to display the total neuropathology. However, animal models – because of the species barrier – have a number of additional limitations for a successful drug development program, which are not always fully appreciated. This chapter discusses (1) differences in drug affinities between human and rodent targets, (2) the absence of key human functional genotypes in rodent models, (3) the intrinsic difference in some neurotransmitter circuits, and (4) the difficulty of simulating the same amount of drug exposure as in the clinical situation. In addition, the problems associated with extrapolating cognitive tests in animal studies with actual performance of treated AD patients on clinical scales are explored. Possible solutions to this dilemma include (1) developing multitarget directed ligands where cholinesterase inhibition is combined with disease modification, (2) a better translation of clinical endophenotypes, (3) capitalizing on drug discovery efforts for cognition in other disease areas, (4) the introduction of realistic polypharmacy in early stages of preclinical tests, and (5) the systematic testing of the face-value of a specific preclinical model/readout combination using marketed drugs with documented clinical effects. Finally, Computational Neuropharmacology, a novel and highly innovative computer modeling approach of interacting brain circuits is introduced. When added to the toolbox of preclinical drug discovery, this approach intends to bridge the difference between preclinical animal models and the clinical situation and reduce the rate of attrition.