By 2050, two-thirds of people with dementia will live in Low- and Middle-Income Countries (LMICs). Efforts to adapt and test multi-modal prevention interventions focusing on lifestyle changes for people at risk of dementia are being developed predominantly in higher income countries, for people with and without cognitive symptoms. However, there is evidence that needs may differ between these groups. We systematically reviewed Randomized Controlled Trials (RCTs) evaluating non-pharmacological interventions in individuals with Mild Cognitive Impairment (MCI) and Subjective Cognitive Decline (SCC) in LMICs. We analysed study quality using the Mixed Methods Assessment Tool (MMAT), meta-analysed and synthesized evidence. We included 25 RCTs, from six countries (most in China, n=17), involving 1304 participants. Across the 15 studies for which there was sufficient data to meta-analyse, we found significant positive effects on cognitive outcomes favouring interventions [1.49 (standardised mean difference, 95% Confidence Interval= 1.06-1.93)]. There was significant publication bias. Using an a priori standard framework, we classified interventions into exercise, multidomain, and arts/creative expression. Group exercise [1.67, 1.24-2.11, n=8]. and multidomain [1.22, 0.22-2.21, n=5] had replicated evidence of effectiveness. There was insufficient data to meta-analyse the creative arts category. We identified one high quality, multi-modal intervention, which combined Chinese mind-body exercises and health education, reporting a positive outcome. The first robust dementia prevention trials are underway in LMICs, evaluating effectiveness of models developed in higher income countries. We propose greater consideration and investment in development of interventions that account for specific LMIC contexts from the outset, so they are acceptable and used by local services.HighlightsThere was evidence that group exercise and multimodal interventions were effective for people with memory concerns in LMICs.We identified publication bias in meta-analyses; the open science agenda is critical to improving care in LMICs and reducing global inequalities.Most included studies were conducted in China, reflecting a need for high-quality evidence from underrepresented regions, including Africa and Latin America.We propose greater investment in developing interventions that account for specific LMIC contexts from the outset, so they are acceptable and used by local services.