At different milestones in the software evolution process, software quality evaluation is a trivial task. This can be used to schedule performance assessment, quality management and project enhancement operations. Two techniques Linear Programming with Multiple Parameters (LPMP) and Quadratic Programming with Multiple Parameters (QPMP) for assessing the quality of software had been employed in the ongoing studies and researches. Several experts conducted research with Support Vector Machine (SVM), Neutral network (NN), C5.0 for quality assessment. These experiments had given poor and low results. In this research, by utilizing corresponding attributes of a multiple datasets, we fine-tuned prediction efficiency. In addition to employing a method of selecting a subgroup of relevant variables and variance matrix for getting greater and better results, we have applied different tests on latest approaches and accomplished good results for other predictive activities. Machine learning (ML) algorithms such as Logistic regression (LR), AdaBoost (AB), Random Decision Forest (RDF), Bagging Classifier (BC) and Classification Tree (CT) are executed on the data to forecast the software functionality, reliability and disclosed the association between the parameters of quality and production. The investigational outcomes proved that the measure of software quality can be well determined and assessed by ML techniques.