The internet-based criminal scheme known as phishing leads to monetary loss and identity theft attacks. New online electronic services, together with payment systems, have created a necessity for accurate detection technologies to combat phishing attacks. The major security risk of phishing attacks requires modern detection systems to address this threat effectively. The proposed research develops a phishing attack identifier based on 1D Convolutional Neural Networks (1D CNN) to achieve effective detection with reduced incorrect alarms. The research methodology begins with obtaining PhishTank dataset information, then proceeds with text tokenization and word embedding before min-max scaling, which leads to classification using the proposed 1D CNN model. Testing found that the proposed model surpassed traditional machine learning models, including Recursive Neural Networks (RNN) and XGBoost, as well as Logistic Regression (LR) according to accuracy together with precision and recall and F1-score metrics. A proposed 1D CNN model surpasses comparable models for detecting phishing attempts by achieving 99.85% accuracy, 99.90% precision, and 99.80% recall and F1score. This result underscores the potential of phishing attack detection, offering a promising tool for cybersecurity applications.