视唱是指通过阅读乐谱并理解其音乐结构，准确地演唱出音乐的技能，旨在帮助学习者提高音乐素养和演唱能力。目前已存在的视唱教育软件缺乏互动乐趣和难度设计，音乐学生在学习视唱技能时面临乐趣缺失和教学计划粗糙的问题。本文提出了一种基于动态难度调整的视唱游戏，本文的主要工作包括：第一，基于开放学习者模型，基于 Unity 引擎设计并开发一款音乐教育游戏《寻星》，游戏实现了自动显示曲谱，以及与人工评分一致性较高的自动评分系统。第二，游戏提出了一种针对音乐教育游戏量身定制的动态难度调整方案。对曲目的结构进行分析，提取出影响视唱表现得核心维度，由正规方程计算玩家水平和这些维度之间的关系，达到预测玩家表现的目的，最后根据玩家行为和水平变化调整难度参数。第三，设计重复测量实验验证视唱评估系统和预测模型的效果，以及设计随机对照实验验证游戏的教育效果。研究实验结果：测试组 1 相较测试组 2，节奏和音准得分变化有较大幅度的提升。测试组 2 相较于对照组，节奏和音准得分变化也有较大幅度的提升。本文的创新点主要体现在：一是数据层面，将动态难度调整融入严肃游戏中，通过曲目分析为音乐教育提供难度依据，构建玩家水平的预测模型，为不同玩家构建自适应难度调节系统；二是内容层面，利用视觉反馈、音乐游戏机制和开放学习者模型提升视唱训练的乐趣，提升玩家的学习动机。

Sight singing is a skill that involves accurately singing music by reading the scoreand understanding its musical structure, aimed at improving musicianship and singingability. However, currently available sight-singing education software lacks interactivefun and difficulty design, resulting in students facing rough lesson plans and a lack offun.This thesis proposes a sight-singing game based on dynamic difficulty adjustment,with the main work being: designing and developing a music education game ”StarSearch” based on the Unity engine, implementing an automatic display of the music scoreand an automatic scoring system? proposing a dynamic difficulty adjustment scheme tailored to music education games? and designing experiments to verify the effectiveness ofthe sight-singing assessment system and prediction model, and the educational effects ofthe game.The results showed that the test1 group had a substantial increase in rhythm and pitchscore changes compared to the test2 group, and the test2 group also showed a significantincrease compared to the control group.The innovations of this thesis are integrating dynamic difficulty adjustment into serious games, constructing a prediction model of player level, and building an adaptivedifficulty adjustment system for different players, while also enhancing the fun of sightsinging training and improving players’ learning motivation through visual feedback, music game mechanics, and open learner models.