Abstract: Music style reflects the total basic characteristic of music work，which is the foundation of music listening，analysis and research.Aiming at the core of music style genre analytical technical：cantus characteristic description and characteristic match，this paper develops the feedforward neural network structure of not-adjacent layer conjunction，gives the machine that the error margin is inverted spread the classification of training the calculate way and carries on the experimental research.The result indicates that the front feedback nerve network structure of not-adjacent layer conjunction has predominant function on identify and splitting velocity on convergence.

Key words: music, style, characteristic, nerve network

摘要： 音乐风格反映了音乐作品的总体基本特征，是音乐欣赏、分析、研究的基础。针对音乐风格流派分析技术的核心——旋律特征描述和特征匹配，发展了非毗邻层连接的前馈神经网络结构，给出了误差反传训练算法的分类器，并进行了实验研究。结果表明，非毗邻层连接的前馈神经网络结构有优越的识别性能和极快的收敛速度。

关键词: 音乐, 风格, 特征, 神经网络