The widely used word vector representation is incapable of fully representing the specialized texts and phrases in sphere of highly specialized chemical industry,which were quite professional and complex,resulting in the low accuracy of classification.Here,we propose a text classification model incorporating multi-granularity dynamic semantic representation.First,the adversarial perturbation was introduced into the word embedding layer of the model to enhance the ability of dynamic word vectors to represent the semantics.Then the word vector weights were redistributed by a multi-headed attention mechanism to obtain a better textual representation of key semantic information.Finally,text representations of different granularities were extracted through the proposed multi-scale residual shrinkage deep pyramidal convolutional neural network (MSRS-DPCNN) and hybrid attention capsule bidirectional LSTM (HAC-BiLSTM) network model,which were then fused for classification.The experimental results showed that the proposed model achieved an F1-score up to 84.62％ on the chemical domain text dataset when using different word vector representations,an improvement of 0.38-5.58 percentage points compared with existing models.The model also had pretty good generalization performance on the publicly available Chinese dataset THUCNews and the Tan Songbo hotel review dataset ChnSentiCorp.