There is an increasing number of cognition-enhancing drugs for Alzheimer's Disease (AD) and, consequently, drug trials represent a growing field of interest in research. As memory dysfunction is generally the first and most severe cognitive impairment in AD, the choice of memory testing to be used in these studies is of great importance. It should reflect an understanding of memory systems being assessed with neuropsychological tests and the fact that some tests can be more appropriate than others to show benefit with certain classes of cognition-enhancing drugs. Severe deterioration of episodic and semantic memory occurs very early in the AD process while working memory shows a gradual deterioration over time. Some aspects of working and implicit memory can be spared in the mild to moderate stages of AD. Tests of working, episodic, semantic and implicit memory are used as outcomes in trials with acetylcholinesterase inhibitors, drugs with other neurotransmitter strategies, metabolic enhancers and drugs which may impact upon a variety of CNS processes. The clinical scales and observational measures are largely used in trials of cognition-enhancing drugs for AD (46.66% of all the studies reviewed). The Digit Span test, the Rey Auditory Verbal Learning Test, the Buschke Selective Reminding Test and the verbal fluency tasks are the most sensitive memory tests, whereas the most sensitive scales are the Sandoz Clinical Assessment-Geriatric, the Gottfried-Bräne-Steel scale and the Blessed Dementia Scale. Finally, we suggest that future investigations should use sensitive memory tests, together with behavioural and psychiatric scales, rather than general observational evaluations.