Sound event detection involves automatically recognizing different sounds. One of the main challenges in this area is that sound data is sequential, which makes it difficult to capture the relevant features and train appropriate models. One of the primary challenges in this field is that sound data is sequential, which makes it challenging to capture the relevant features and train suitable models. Our research was conducted on sound event classification on a dataset from Kaggle related to smart homes that had 18,873 audio files sorted into 41 different classes. The model was implemented with a Convolutional Neural Network (CNN) on a selected dataset for better classification. This CNN model can analyze various features of sound data, such as statistical features (e.g., mean, minimum, maximum) and content-based features like bandwidth and frequency. These features are important for accurate sound event detection. Our model was trained using a CNN architecture that utilized ReLU and softmax activation functions and had a learning rate of 0.001, as well as a dropout rate of 0.40. The model was trained for 100 epochs and reached a precision of 81% on a large dataset.