The Forensic investigation often involves the classification of digital artifacts into different categories for analysis. Traditional classification methods rely heavily on manual intervention and predefined rules, leading to limited scalability and adaptability. This paper proposes a novel three-tier forensic classification framework leveraging deep learning techniques to automate and improve the accuracy of classification tasks. The first tier of the framework involves data preprocessing and feature extraction using deep neural networks (DNNs) to transform raw digital artifacts into meaningful representations. The second tier employs convolutional neural networks (CNNs) for image-based artifact classification, capturing spatial dependencies and patterns within the data. The third tier utilizes recurrent neural networks (RNNs) for sequential data, such as text or network traffic, to capture temporal dependencies and context. Experimental results on a real-world forensic dataset demonstrate the effectiveness of the proposed framework, achieving state-of-the-art performance in artifact classification. The framework's modular design allows for easy integration of new artifact types and adaptation to evolving forensic scenarios. Overall, this framework presents a promising approach to enhance the efficiency and accuracy of digital forensic investigations through deep learning technology.