The popularity of fingerprints as the means of an individual's identification has increased, and has resulted in a drastic increase in the size of fingerprint datasets. Every fingerprint can be classified based on their patterns. This can be done with the help of a classification system, which also helps to reduce the search and space complexity of the identification algorithm. Random Forest (RF) algorithm is an ensemble-learning method proved to be very effective in the field of bioinformatics. The system proposed in this paper applied the RF algorithm to help classify the fingerprints, and found that it produces highly accurate results in reasonable time. In addition, different features of the fingerprints such as number of ridge-endings and bifurcations, number of cores and deltas and average inter-ridge distance, were used to build a gender prediction model, which was found to be reasonably accurate.