Abstarct: The management and administration of countries is baxsed on accurate and correct laws. Establishing appropriate and timely laws can lead to the progress and excellence of a country, and extracting useful knowledge from the large volume of these laws can improve the performance of organizations and legislative institutions. Each law is a textual phrase that is added to the set of existing laws after going through a process with the approval of the parliament. In the review of each new law, among the set of existing laws, related laws are extracted and analyzed. In the legislative process in the Islamic Council, first, a proposal or bill that is placed on the agenda of the parliament is examined and extracted by the experts of the laws of the parliament. In this treatise, a new solution for extracting related rules for a discussed phrase from among the set of existing rules using semantic communication and deep learning techniques baxsed on Brett's model and baxsed on Transformer architecture is presented, and in this method, sentences or text paragraphs are encoded in a fixed-length vector (dense vector space) and then those vectors are used to evaluate and score the semantic relevance of sentences with the cosine distance measurement scale. And in this way, the model recognizes which words are related and selects them. In this way, the machine can understand the meaning and meaning of the sentences by using the Burt model coding method, taking into account the position of the entities used in the sentences. to understand and to discover the semantic relationships of the documents and to calculate the degree of connection of their documents with a subject and to recognize their semantic similarity. The results obtained from a small sample of test data (about 100 draft bills and 150 related law clauses) show the correctness and accuracy of the proposed method in detecting the semantic relationships of legal documents related to the Islamic Council and show that the accuracy and accuracy of the performance is above 92%. It gives and accelerates and improves the quality of legislation in terms of content, costs and time, and also seeks people's satisfaction due to accuracy in legislation and reduction of costs and time.