Methodology used for the development of anti-Alzheimer's disease (AD) drugs raises specific problems which are rarely examined in the literature. While the general development scheme is similar to that required for most drugs, some specific aspects must be analyzed, highly dominated by the dual goal of pharmacology, i.e., to obtain both symptomatic and etiopathogenic drugs. During preclinical studies, aged or lesioned animals are mainly useful for symptomatic drugs, whereas transgenic models or neurodegeneration-induced techniques would probably lead to etiopathogenic drugs potentially slowing down the process of AD. The first administrations of a new compound to human beings raise the question of the activity measurement techniques. Psychometry remains the most informative procedure to detect and analyze the activity of the drugs on the different components of cognition. Electrophysiology and neuroimaging need some complementary studies before they can be proposed as surrogate criteria in phase III trials. At this stage of development, American and the recently published European guidelines are of great help while insisting on long-term (6 months) placebo controlled trials with the use of the triple efficacy criterion: an objective cognition scale, a global assessment, and the opinion of the caregiver. In the long term, pharmacoepidemiology and pharmacoeconomy will have to confirm the rationale of this recent progress in the methodology of anti-AD drug development.