Now a days, Internet activities are growing in exponential rate so are the criminal activities, with the growth of internet usage. Internet is also a source of malicious web pages. Automatic Malicious URL identification resulted a relative novel and sensitive security challenging area. The area would aim in aiding the users to overcome security threats due to the presence of malicious webpage's resulting in a better network security. The present study makes attempt in assessing and identifying malicious websites, a malicious identification model is proposed using deep learning ideas. The present work uses the URL and HTML based features to identify malicious websites. PCA is applied to reduce features, dominating features are identified. It is found that dominating features play vital role in segregating the URLs into malicious and non-malicious. Dataset from PhishTank and Alexa is used in this study. Seven Layer Neural Network has shown significant improvements resulting in accuracy of 94%. The proposed work gave true-positive rate 95.51 and Falsemalicious rate 9.51.