The growth of the Semantic Web has seen a rapid increase in the amount of Resource Description Framework (RDF) data. Meanwhile, the demand for access to RDF data without detailed knowledge of RDF query languages is increasing. In this study, an approach enabling keyword-based semantic information query over RDF data is proposed. The approach sets up a keyword-inverted index and a relation index based on the r-radius+ graph and searches the connecting nodes to provide an answer for keyword query. Moreover, the approach uses an improved scoring function based on textual relevancy and relation popularity and supports top-k queries. Experimental results show that the proposed approach can achieve good query performance.