Data mining is the procedure of hauling out enviable information or remarkable patterns from presented databases for precise purposes. The effectiveness of the rules produced depends on the support threshold, which consecutively involve decisions finished employing these rules. Nearly all of the earlier strategies put a distinct minimum support threshold for all the items or item sets. But in genuine applications, diverse items might have diverse criteria to review its significance. The support necessities ought to then differ with diverse items. The existing work presented a structure for discovery of patterns based on propositional logic which evaluate the coherent rules (i.e., knowledge discovery). The discovery of association rules openly from the logical rules with no minimum support threshold is evaluated. Nevertheless it processes on distinct level of propositional logic, where hierarchical schemes of the knowledge domain cannot be derived. To enhance the pattern discovery process, the proposed work extends the pattern discovery process with coherent rule generation framework in terms of multi-level hierarchical property propositions. The multi-level coherent rule structure produce rules coming from diverse levels and determine highest recurrent item sets at inferior level. The propositional logic process formed the multilevel connection rules from logical rules and utilizes bottom-up progressive extending technique. The bottom up progressive method develops the effectiveness of rules being produced devoid of minimum support threshold. Experimentation are carried out using real data set to assess multilevel association rules capably using concept hierarchies, which describes a series of mappings from a position of low level concepts to advanced level. The presentation of rule creation is measured up to that of the presented single level coherent rule miners.