It is difficult task to select optimal web service from the list of functionally equivalent  web services. In case of Internet services, the presence of low-performance servers, high latency or  poor service quality directly affected on lost sales, user frustration and customers lost. We propose a novel method for Quality of Service metrification based on Hidden Markov Models. HMM suggest then optimal path for user request. The HMM technique can be used to measure and predict the state or behavior of Web Services in terms of response time, and can be used to rank services quantitatively instead of just qualitatively. We demonstrate the feasibility and usefulness of our methodology by doing experiments on real world data sets. The results have shown how our proposed method can help the user to automatically select the most predictable Web Service taking into account several metrics, among them, system predictability and response time variability.