With the advent of the era of big data, people face Web information overload problem. In order to efficiently and accurately obtain information on the network, Web information search technology has been developed rapidly. First in this paper, the working principle of the keyword search engines that are the traditional Web information search technology is reviewed, and then from the recall ratio and precision, reasoning search, information in a timely manner and the user experience etc, we analyze its existence disadvantages and Improvement strategy. Finally we present the train of thought and key technologies of a new Web information search that is based on networked automatic reasoning. This paper provides a train of thought for building automatic reasoning system which is adapted to the network of open, distributional and dynamic, and also puts forward a new direction for the development of Web information search technology and the architecture of next generation of search engine.