Plastic waste in oceans is a global environmental concern which endangers the marine ecosystem, biodiversity, and human livelihood. Each year, millions of tons of plastic waste are reaching the oceans, leading to ocean contamination. Hence, detecting these pollutants is extremely essential for effective cleaning and prevention. Conventional methods of detecting ocean plastic litter are less efficient, expensive, and mostly limited in scope. Therefore, innovative technologies like AI-driven detection systems must be applied for oceanic plastic waste detection. In this paper, oceanic plastic waste is detected using computer vision by applying 4-layer Convolutional Neural Network model for binary image classification tasks, leveraging a multi-layered architecture to extract and process spatial features from input images. The network is designed to handle 224 × 224 RGB images and comprises convolutional layers for feature extraction, max-pooling layers for dimensionality reduction, and fully connected layers for classification. The inclusion of dropout regularization mitigates overfitting, ensuring robust performance. The final softmax layer outputs probabilities for the two target classes, namely, plastic and non-plastic images. This architecture demonstrates efficiency in capturing hierarchical image features, making it suitable for real-world image classification applications. The trained model is integrated into a real-time system for detecting oceanic plastic pollution. This system aims to detect and categorize plastic waste effectively, thereby safeguarding the health of our planet’s oceans and promoting sustainability.