© 2015 IOS Press and the authors. All rights reserved. Background: The detection of early Alzheimer's disease (AD) can rely on subjective and informant reports of cognitive impairment. However, relationships between subjective cognitive impairment, objectively measured cognitive function, and amyloid-β (Aβ) biomarkers remain unclear. Objective: To determine the extent to which impairment or decline in subjective and informant rated cognitive impairment was associated with memory in healthy older adults with high Aβ. Methods: Healthy older adults (n = 289) enrolled in the Australian Imaging, Biomarkers and Lifestyle (AIBL) study were studied at baseline. Pittsburgh Compound B was used to determine Aβ status at baseline. At baseline and 18 months assessments, subjective memory impairment was assessed using the Memory Complaint Questionnaire and the Short Form of the Informant Questionnaire on Cognitive Decline in the Elderly. Cognition was measured using the Cogstate Brief Battery. Results: At baseline, there were no differences between low and high Aβ groups in subjective or informant-rated cognitive impairment, depressive and anxiety symptoms, or cognitive function. Longitudinal analyses showed moderate decline in learning and working memory over the 18 months in the high Aβ group. However there was no change over time in subjective or informant-rated cognitive impairment, depressive and anxiety symptoms, or cognition in either Aβ group. Conclusions: Although healthy older adults with high Aβ levels show decline in learning and working memory over 18 months, subjective or informant ratings of cognitive impairment do not change over the same period suggesting subjective cognitive impairment may have limited utility for the very early identification of AD.