Clinical differentiation between Alzheimer's disease (AD) and depression is often difficult due to symptom overlap and similar clinical presentation. Concise and accurate diagnostic tests have been of interest for many years. Furthermore, with the continued growth of the Cattell-Horn-Carroll (CHC) theory, there has been an emergence of measures such as the Woodcock-Johnson-III Tests of Cognitive Abilities (WJ-III COG), which are being more commonly used in clinical practice yet have not been fully evaluated in terms of their efficacy in various domains of clinical practice. This study investigated the predictive and diagnostic properties of the WJ-III COG as a CHC-based test in differentiating between AD and depression in the elderly population. A discriminative function used in the study was able to correctly classify 89.02% of cases using six areas of the CHC framework. The variables that had the highest predictive weights were: long-term retrieval, fluid reasoning, processing speed, and working memory. The main implication is that a theory-based approach is crucial in the support of differential diagnosis and in decreasing the length of assessment for elderly populations.