The purpose of this paper is to develop a memory-based prediction system of learners′ score in the assembly process of a manufacturing class, and to evaluate the precision. Thinking scale and an achievement test in the assembly were conducted on 274 junior high school students, who were divided into two groups : learning data (n = 214) and inspection data (n = 60) . We designed the memory-based prediction system, which accumulated learning data. The result of the precision evaluation with inspection data showed that the correlation between measured and predicted scores was 0.51 and the absolute value of errors was under 0.1 in 81.7% of the total predicted scores.