Recognizing and extracting name entities like person names, location names, date and time from an electronic text is very useful for text mining tasks. Named entity recognition is a vital requirement in resolving problems in modern fields like question answering, abstracting systems, information retrieval, information extraction, machine translation, video interpreting and semantic web searching. In recent years many researches in named entity recognition task have been lead to very good results in English and other European languages; whereas the results are not convincing in other languages like Arabic, Persian and many of South Asian languages. One of the most necessary and problematic subtasks of named entity recognition is person name extracting. In this article we have introduced a system for person name extraction in Arabic religious texts using proposed “Proper Name candidate injection” concept in a conditional random fields model. Also we have created a corpus from ancient Arabic religious texts. Experiments have shown that very hight efficient results have been obtained based on this approach.