Movie recommendation systems are a little challenging and temporary in not delivering accurate and personalized movie suggestions. This paper delves deeply with the integration of Neural Network models (Graph Neural Networks (GNN), Feedforward Neural Networks (FNN), and Recurrent Neural Networks (RNN) and Convolution Neural Networks (CNN) with a focus on item-based Collaborative filtering (CF) to give enhanced movie recommendations. It also encourages users (customers) to experiment with new genres and take joy in a customized movie-watching experience. This data is also an advantage for content recommenders on entertainment platforms and online streaming services. We have a designed personalized dataset with customized attributes for recommendation model. To give a solid analysis, the dataset is divided into parts for testing and training, which provides the user with MSE for the top five recommendations by leveraging the strengths of these diverse neural network models. The core focus of this hybrid model is to enhance the precision and personalization of movie recommendations generated for users.