Movies represent one of the most important media that combine narrative components with visual and auditory elements to entertain or inspire viewers. The Motion Picture Association of America (MPAA) is a prominent rating system that plays a great role for audience to select an appropriate film. This rating system is important because it helps parents filter the content to protect their children from unsuitable movies. It also assists audiences in their own choice. Traditionally, MPAA ratings are assigned by reviewers manually and this, in turn, leads to inconsistency, subjectivity, and time-consuming. This research introduces an automated method for predicting MPAA rating by using the scripts feature with the emotion feature that aiming to enhance classification accuracy and provide a more reliable assessment of age appropriateness. Scripts are preprocessed and converted into term frequency-invert document frequency (TF-IDF) vectors to capture significant linguistic patterns. Moreover, emotional features are extracted from movie scripts using transformer-based models due to their contextual understanding capabilities. These features are integrated to form a multi-class output. In this study, the LightGBM algorithm is applied as a gradient boosting technique. Experimental findings indicate that combining emotion features alongside textual representations enhances prediction accuracy in comparison to the use of scripts alone. The model achieves 84.2% and 84.6 for the weighted F1-score and the accuracy metric, respectively. This supports the effectiveness of the proposed model in predicting MPAA ratings from both movie scripts and emotional features.