OBJECTIVES: To determine the prevalence and predictors of cognitive impairment among older residents of nursing homes in the state of Penang, Malaysia. METHODS: This multi-centred cross-sectional study was conducted in two phases over a six-month period at four non-governmental nursing facilities in Penang, Malaysia. Older residents (≥65 years old) who were taking at least one medication and had not been previously diagnosed with dementia were included. Demographic and clinical data were collected through direct interviews and the review of medical records. Cognitive screening was performed using the Mini-Cog at baseline. Health Related Quality of Life (HRQOL) was assessed at baseline and after a three-month interval using the EuroQol (EQ-5D) and EQ Visual Analogue Scale (VAS). RESULTS: : The study included 211 residents with mean (SD) age of 77.7 (7.0) years and majority female residents, 128 (60.7%). Mini-Cog identified 129 (61%) residents with cognitive impairment. The three most common co-morbidities among the sample population were hypertension (71.1%), diabetes mellitus (27.1%) and cerebral vascular disease (12.3%).  Logistic regression analysis shown that longer formal education of more than six years (OR = 2.81, p = 0.007, 95% CI [0.167-0.757]) and inability to self-administer medications (OR = 3.29, p < 0.001, 95% CI [0.156-0.594]) significantly predicted cognitive impairment based on Mini-Cog score ≤ 2. However, Mini-Cog was not significantly associated with changes in HRQOL at three months follow-up.  CONCLUSIONS: We found a high prevalence of possible cognitive impairment among nursing home residence who had not been previously diagnosed, especially those who were unable to self-administer their medications. The length of formal education can be deceptive. Therefore, more vigilant screening should be performed for early detection and proper intervention of dementia although HRQOL is not affected by poor scoring in Mini-Cog.