Random Forest (RF) also known as Attribute Bagging, is one of the most popular ensemble supervised learning algorithms in machine learning. It creates random tree in its each iteration by randomly selected features. Random Forest can use any decision tree induction methods for selecting best node of the tree. It’s a robust algorithm and requires little prior knowledge for model building. Random forest can be used for multi-class imbalanced classification. In this paper, we have amalgamated Random Forest with Z-score, also known as a standard score, for classification and regression tasks. The proposed idea is to select the features randomly and then apply Z-score on the randomly selected features to build random tree. Z-score indicates how many standard deviations a data point is away from the mean of a distribution. We have used four benchmark datasets (two for classification: prediction of diabetes and heart disease and two for regression: air quality prediction and salary prediction) from Kaggle Dataset. The effectiveness of the proposed Random Forest technique is validated by the accuracy, precision, recall, F1-score for classification, and RMSE (Root Mean Squared Error), and R \(^2\) value for regression.