The long-term average rainfall for a given period (month, season or year) scarcely indicates reliability because rainfall in low latitudes varies significantly from one year to the other. The less variable rainfall is, the more reliable it is, as the index of variability measures the likelihood of repetition in the mean amount of rainfall. This study bridges some gaps in related studies in the Bamenda Highlands, such as a study assessing climate change impacts on food security, where the Rainfall Anomaly Index was used and another study on the impact of rainfall on maize production using the Standardized Precipitation Index. The objectives of this study are to assess rainfall reliability and establish the impact of rainfall reliability on maize production. Rainfall data were collected from 15 stations, while maize output was collected for four localities. Results revealed that rainfall is still reliable for 13 stations, with a coefficient of variations of 9.62 to 18.54 %, while Ndop and Ndawara recorded unreliable rainfall of 23.14 % and 30.97 % respectively. Rainfall reliability was complemented by the Standardized Precipitation Index, which showed that only 53.45 % of rainfall episodes were normal to sustain maize production, while 46.55 % were events of rainfall deficits. Maize production has been decreasing in Ndu, Oku and Nkum while increasing in Ndop. These findings reflect the realities of other tropical mountainous regions in the world. Faced with future climatic uncertainties, farmers should embrace agroecological practices, climate-smart agriculture, conservation agriculture, Nature-based Solutions, Ecosystem-based Adaptation and diversification of production systems and livelihood sources to ensure food security.