Working Memory Binding (WMB) entails the integration of multiple sources of information to form and temporarily store unique representations. Information can be processed through either one (i.e., Unimodal WMB) or separate sensory modalities (i.e., Crossmodal WMB). Objective: In this study, we investigated whether Crossmodal WMB is differentially affected by normal or pathological ageing compared to Unimodal WMB. Methods: Experiment 1: 26 older and 26 younger adults recalled the target feature matching the test probe to complete a previously displayed colour-shape binding (visually presented in the Unimodal condition; auditorily and visually presented in the Crossmodal condition). Experiment 2: 35 older and 35 younger adults undertook the same paradigm while carrying out articulatory suppression to limit verbal recoding. Experiment 3: 24 Alzheimer’s Disease (AD) patients and two groups of 24 healthy matched controls (tested respectively with the same and an increased memory load compared to the patients) were recruited to perform a similar task. Results: Results show no age-related additional cost in Crossmodal WMB in respect to Unimodal WMB. AD patients had poor attainment in both WMB tasks regardless of specific binding condition. Conclusion: These findings provide evidence identifying WMB per se to be impaired in AD, regardless of the type of to-be-bound material. This supports the view that WMB is a suitable cognitive marker for AD.