受测者在测试过程中存在猜测和失误的可能，然而现有的诊断模型并未考虑这两个因素，因此无法在这种情况下对受测者的知识状态进行有效诊断。本文提出了一种知识空间理论中基于猜测概率和失误概率的学习诊断模型，当受测者出现猜测或失误时，该模型仍可对受测者的知识状态进行有效的诊断。模型依据试题的猜测概率和失误概率对答题状态进行转移，以诊断受测者的真实知识状态，并重点分析了这两种概率的计算方法。该模型在实用性、诊断的准确性等方面具有较好的效果。

There exists the possibility of guess and misplay when testing, however, the existing diagnosis models do not take these two factors into account, so it is difficult to diagnose the knowledge states of students effectively. Based on the guess probability and the misplay probability in the knowledge space theory, a new model is proposed which could effectively diagnose the knowledge states of students in that case. To diagnose the real knowledge states of the students, the model transfers the answer states according to the probabilities of guess and misplay of each question, and focuses on the analysis of the calculation of the two probabilities. The model has a good effect on practicability and the veracity of diagnosis.