Compressive strength of masonry in historical buildings was evaluated based on Bayesian method. The limited information and divergent measured information were incorporated with reliable prior information, and information of direct measured strength samples and information of indirect measured strength samples were incorporated through an appropriate likelihood function. Probability density functions (PDFs) for compressive strength of brick, compressive strength of mortar, and compressive strength evaluation error of masonry were established. Compressive strength of masonry was inferred and its uncertainty was quantitatively evaluated based on the PDFs. The proposed method is suitable for compressive strength inference for masonry in historical buildings under limited measured information.