It is essential to use search engines to get the needed information. A search engine uses result matching to match the user’s query with appropriate web pages. Users see the search results in a certain order based on how they are ranked. A website or web page may be made better using search engine optimization (SEO), which will increase the amount of organic traffic it receives from search engines. If we can’t manage efficient SEO techniques to rank at the top of organic search results, a lot of money will be spent on sponsored adverts for certain keywords. The process of building rank estimate algorithms for search engine results pages (SERP) or applying data analysis to find the best SEO tactics has been employed in several research projects. The datasets of existing studies were undiversified since they only included web pages from one or a small number of categories. This research will improve rank estimation algorithms by using multi-category web pages in the training datasets and will provide demonstrations of improvement on SERP rank estimation algorithms for English web pages. Since Google receives more than 90% of all internet search submissions, scraping will be used there. For the chosen web pages, a collection of on-page SEO variables will be retrieved. The methodology starts with choosing a set of search terms and scraping search engines, then crawling SERP web pages to extract certain SEO criteria from the contents of web pages, and lastly getting to data preprocessing. Various machine learning techniques were used to compare performance and choose the optimal approach. The main finding of research is the enhancement of SERP rank estimation by more than 25% on performance with the proposed dataset improvements for building models.