In the ever-evolving realm in the field of cybersecurity is dynamic in the sense that new developments are being recorded frequently; therefore, zero-day malware poses a lot of problems for defenders of these digital ecosystems. More known by the unnamed actions that it is capable of performing, zero-day malware configures itself at the cutting edge of risks where it capitalises on an opportunity before the weakness is identified. The research proposed here is about detecting Android mal-wares using machine learning (ML) and as a part of it, how ensemble learning can be used as a metamodel that can be implemented through bagging and stacking techniques using ML. This model fusion increases the accuracy of hybrid ensembled models up to 98.00% with high levels of system reliability in avoiding both false negative and false positive results. The authors assess their models using actual data sets in this case in a bid to posit that ensemble learning has a better performance as an individual classifier more so in android mal-ware detection as it is well suited to the new emerging threats as compared to one classifier. Based on the literature review, there is scope for increasing the level of security of portable devices on Android platforms by further developing the use of ensemble learning for android malware detection; these findings may therefore serve as a useful guide for researchers seeking to increase safety of such mobiles.