Resolution-based automated reasoning is one of most important research directions in AI, semantic method is one of the most important reform methods for resolution principle, in semantic resolution method, it utilize the technology that restraining the type of clauses and the order of literals participated in resolution procedure to reduce the redundant clauses, and can improve the efficiency of reasoning, α - resolution principle on lattice-valued logic based on lattice implication algebra provide a alternative tool to handle the automated reasoning problem with uncomparability and fuzziness information. it can refutably prove the unsatisfiability of logical formulae in lattice-valued logic system. Firstly, this paper discussed the property of one class of generalized clause set on lattice-valued propositional logic LP(X), this generalized clause set can be divided into two non-empty sets, the semantic resolution method on it is investigated and sound theorem and weak complete theorem of this semantic resolution method were proved.