In the world of professionals, figuring out how much someone should get paid can be tricky. This work goes into the complex world of salary prediction. This study presents a comprehensive approach to salary prediction using machine learning techniques, incorporating extensive data preprocessing and advanced model optimization. The dataset, derived from the Stack Overflow Developer Survey, includes variables such as Country, Age, Remote Work status, Education Level, Years of Professional Coding Experience, and Salary. Initial preprocessing involved filtering the dataset to include only full-time employed individuals, mapping countries based on frequency, randomizing age within specified ranges, and standardizing experience and education levels. Outlier salaries are adjusted to mitigate the impact of extreme values. Linear Regression, Decision Tree, and Random Forest are employed here to predict salaries. The Random Forest model demonstrated the best performance, with a Mean Absolute Error (MAE) of 17,166, a Root Mean Squared Error (RMSE) of 22,031, and an Accuracy of 81%. To further enhance the model performance, a unique optimization approach is applied, involving extensive hyperparameter tuning and cross-validation. This study underscores the importance of thorough data preprocessing and hyperparameter optimization in improving the accuracy of salary prediction models, providing valuable insights for both researchers and practitioners in the field of machine learning. This research bridges the gap between traditional salary models and the evolving landscape of employment, providing a valuable tool for organizations seeking precise and data-driven salary estimations.