Past decades, the natural disasters such as flash flood and landslides in the Cameron Highlands have caused significant damage to public infrastructure and posed serious risks to human health. This study aims to investigate projected changes in precipitation under various future scenarios and spatiotemporal scales. Specifically, it focuses on statistically downscaling the CANESM5 model and analysing rainfall pattern changes for the period 2015–2100 under the SSP2-4.5 and SSP5-8.5 scenarios through spatial analysis. By comparing historical observed and simulated data, the study confirms that the model effectively represents rainfall pattern across the projection period. Results from the localized climate model (2015–2100) indicate that precipitation trends at Stations 1 (p1-Kg. Raja township) and 2 (p2-Bertam valley) are similar under both scenarios until 2063, after which they diverge. In contrast, Station 3 (p3-Brinchang township) shows markedly higher precipitation changes. Notably, under SSP5-8.5 scenario, the rainfall anomalies shift from negative to positive around 2049, with this scenario having a stronger influence on precipitation patterns than SSP2-4.5 and being less impacted by terrain effects. Spatial analysis further reveals a correlation between elevation and precipitation, indicating that higher altitudes are associated with increased rainfall.