Data mining has emerged as one of the domains in the field of research. It is an analytic process designed to explore, in search for consistent patterns and systematic relationships between variables in a dataset. In data mining, patterns in huge data are analyzed in order to extract useful information or knowledge. Discovering hidden information from historical data is among its important tasks and one of its ultimate goal is prediction. Prior to the data mining process, data cleaning and preprocessing is performed. In this paper, the Principal Component Analysis (PCA) was utilized to preprocess the KDD Cup 99 dataset. The goal is to address data dimensionality, by reducing noise and remove redundancy, to generate the useful feature subset that has high influence in predicting network intrusions and reduce computational time. The experiment used the WEKA software, specifically the J4.8, RandomTree and RandomForest decision tree algorithms that are capable of detecting intrusions. The algorithms was trained using ten (10) fold cross validation and the generated model was applied, tested. The results were compared between the original over the reduced dataset. Analysis of the results revealed improvements in detecting network intrusions in contrast the original dataset. This can be attributed to the PCA as a feature reduction mechanism applied as a preprocessing technique. Similar studies may be conducted using other classification algorithms and integrating other data mining techniques.