The blocking job shop problem (BJSSP) is a variant of the classical job shop problem, where a job that has been completed on one machine cannot be delivered to the next machine until that one becomes free or available, thus generating the blocking of the current machine. Many real-world problems have been modeled as a BJSSP. This problem is considered Strongly NP-Hard. Existing literature shows that some methods proposed do not guarantee that a feasible solution is generated during the search process, so it has to be reconstructed. In this work, we present an Estimation Distribution Algorithm that operates strictly within the BJSSP feasible solution space. The results show that our method improves some of the previous best known results in the current state of art.