Handling the missing data is very important during the preprocessing stage. And how missing is managed plays a crucial role in enhancing the performance of the machine learning models. In order to get reliable results, we have to handle the missing data carefully. There are studies conducted on different imputation techniques but there is lack of side-by-side comparison. To address this issue, the study examines nine imputation techniques with four classification models such as, Logistic regression, K-nearest Neighbour, Support Vector Machine and Random Forest. and hyperparameter tuning was used. The research shows how each imputation technique performs. Imputation techniques like, multiple imputation by chained equations (MICE) and even simple imputers showed balance between model accuracy and data missingness. On the other hand, techniques like Complete case studies (CCA) and random imputations performed poorly. Especially with simple models. Steady and reliable outcomes received with the help of K-Nearest Neighbours (KNN) imputation. Random Forest was the most robust against the distortions introduced by imputation. Altogether, the study proved that selecting an imputation method should depend on the nature of missing data and also on the complexity and sensitivity of the machine learning model.