This work is focused on predicting melt pool dimensions for multi-objective optimization of microstructural features like porosity, grain size, and mechanical properties of laser powder bed fusion (LPBF) processed Inconel 718 (IN718). The methodology involves developing a dataset comprising a range of input parameters, including laser power, scan speed, hatch spacing, and their corresponding output parameters, such as melt pool width, depth, and grain size. The dataset has been generated using ANSYS Additive Print simulations. This is followed by training machine learning (ML) models on the dataset to predict the microstructural features. The model's accuracy is validated by simulations and experimental data in the literature. This model serves as a guide to determine the optimal process parameters required to achieve the desired porosity and grain size. The proposed ML framework optimizes process parameters for LPBF-manufactured Inconel 718. This has significant implications for defect reduction, material saving, and cost reduction during the product development phase in the industrial environment.