Online safety is frequently and seriously at risk from malicious URLs and websites. Naturally, search engines are the cornerstone of information management. However, our users are now seriously at risk due to the widespread presence of bogus websites on search engines. The majority of methods used today to identify rogue websites focus on a specific attack. Online safety is frequently and seriously at risk from malicious URLs and websites. Naturally, search engines are the cornerstone of information management. However, our users are seriously at risk due to the rise of bogus websites on search engines. The majority of methods used today to identify rogue websites focus on a specific attack. However, a lot of websites remain unaffected by the widely accessible blacklist-based browser add-ons. Any data leaving the client side must be properly disguised, as the server cannot infer any meaningful information from the masked data. Here, the recommended initial Privacy-Preserving Safe Browsing (PPSB) service is given. Robust security assurances are given, which the existing SB services do not offer. The suggested method uses blacklist storage to identify malicious URL access. SVM classification was used to classify the user-provided input URL. SVM is a class of machine learning algorithms that reliably determines the safety or riskiness of a URL. Specifically, it retains the ability to identify malicious URLs while protecting the user's privacy, browsing history, and proprietary data of the blacklist provider (the list of dangerous URLs). This paper presented a technique that encrypts critical data to safeguard user privacy from outside analysts and service providers. Furthermore, completely supports the functions of chosen aggregates for analysing user behaviour online and guaranteeing differential privacy. The AES encryption method is used to protect user behaviour data online.