Older adults with poor sleep tend to show a discrepancy between objective and self-reported sleep parameters, which can trigger a vicious cycle that worsens their sleep complaints. Cognitive-behavioural therapy can reduce this discrepancy, but alternative behavioural therapies remain untested. The present exploratory study aimed to investigate the effects of mindfulness-based therapy for insomnia (MBTI) on reducing sleep discrepancies in comparison with a sleep hygiene, education, and exercise programme (SHEEP). Older adults were randomly allocated into the mindfulness-based therapy for insomnia group (n = 55) or the sleep hygiene, education, and exercise programme group (n = 58). Subjective and objective sleep parameters were measured using sleep diaries, polysomnography (PSG), and actigraphy. Sleep discrepancies were calculated using the Bland-Altman method for sleep onset latency (SOL) and wake after sleep onset (WASO). Additionally, correlations between the change in sleep discrepancies and the change in subjective sleep quality and trait mindfulness were measured within each group. Sleep onset latency discrepancy measured by polysomnography and actigraphy decreased significantly after the MBTI and SHEEP interventions. In contrast, there was no significant change in wake after sleep onset discrepancy in either group. The change in sleep onset latency discrepancy was correlated with the change in insomnia symptoms and objectively measured trait mindfulness. Mindfulness-based therapy for insomnia was effective in reducing sleep onset latency discrepancies and improving sleep perception in older adults with sleep disturbances, which in turn drove an improvement in sleep quality and insomnia symptoms. Increases in trait mindfulness may have been an important mechanism in improving sleep perception in the mindfulness-based therapy for insomnia group.