Internet of Things (IoT) has achieved massive development on last ten years, in network communications. However, IoT produced sensitive data through the Internet, so security become critical due to the increase in attacks. To address these issues, a Deep Convolutional Neural Network (DCNN) for IoT-based effective anomaly detection technique is proposed. The data is then collected and stored using the BoT-IoT dataset. Two processing phases are used to the proposed model: DCNN-based anomaly detection and data preparation and collection. The data is processed by Deep Learning (DL)-based anomaly detection method, which then analyzes for unusual traffic to protect the IoT network. Then the process is performed in the DCNN stage after the data preparation phase. When compared to other models like CNN, DNN, and SEMI-Gated Recurrent Unit (SEMI-GRU), the proposed algorithm performed better in terms of accuracy of 99.65%, precision of 99.54%, F1-measure of 99.56%, and recall of 98.65%, respectively.