Autism spectrum disorder is a complex, lifelong developmental disability where the affected people show repetitive behavior and faces abnormal communication challenges. The goal of this work is to propose an enhanced machine learning model that detects autism more accurately. Hence, we collected ASD datasets of toddler, child, adolescent, and adult from kaggle and UCI machine learning repository. The correlation among individual features was scrutinized and eliminated highly co-linear features in these datasets. Then, feature transformation methods including standardization and normalization were applied in these datasets. Different classifiers like artificial neural network, recurrent neural network, decision tree, extreme learning machine, gradient boost, k nearest neighbor, logistic regression, multilayer perceptron, naïve bayes, random forest, support vector machine, and xgboost were employed in these altered ASD datasets and determined their performances. Logistic regression shows the best result that outperforms other classifiers. This model is useful to extract significant traits and detect autism more precisely.