In order to evaluate the effect of biochemistry blended teaching mode under the condition of uncertain teaching interaction, this paper proposed a fuzzy information computing framework that integrated teaching content mapping, online and offline behavior and multi-index effect judgment. This paper constructed a course dataset containing 186 undergraduates, 14 teaching weeks, 12,640 interaction records, 2,232 test responses, 744 experimental examination records and 368 discussion texts from the learning platform, classroom records and experimental teaching system. The fuzzy learning state is represented by membership coding, and a weighted fuzzy inference engine is designed to model knowledge participation, experiment completion, response timeliness and concept transfer. The evaluation model generates composite effect scores and category labels for different teaching stages. Experimental results show that the classification accuracy of this method is 96.8%, and the Macro-F1 value is 95.9%, which is 7.4 percentage points higher than that of the traditional weighted scoring method. The framework supports data-driven blended teaching evaluation in biochemistry and provides a computable basis for adaptive teaching adjustment, which remains stable across week outputs.
