In this paper, we focus on the semantic relationships acquisition from Chinese web documents motivated by the large requirement of web question answering system in e-Learning. With our scheme, we dwindle in numbers of text to be analyzed and obtain initial sentence-level text in pre-process phase. Then linguistic rules, which are broken down into unambiguous and ambiguous, designed for Chinese phrases are applied to these sentence-level text to extract the synonymy relationship, hyponymy relationship, hypernymy relationship and parataxis relationship. Lastly, candidates are refined using two heuristics. Compared to other previous works, we apply not only strict unambiguous linguistic rules but also loose ambiguous linguistic rules to extract relationships and proposed efficient approach to refine the outputs of these rules. Experiments show that this method can acquire semantic relationships efficiently and effectively.