Abstract: Considering the challenges of insufficient feature fusion and limited classification granularity in electroencephalogram (EEG)-based depression recognition task, a CNN-Transformer encoder-based model for Depressive State Recognition (CTE-DSR) that integrates multidimensional EEG features was proposed. Firstly, one-dimensional convolutional neural network (1D-CNN) was used to extract local spatial sequential features from EEG. Then, Transformer encoder module was introduced to model global correlations between leads. Simultaneously, multi-layer perceptron (MLP) was used to embed depression-related non-sequential features into a high-dimensional space. Finally, the two features were fused and implemented in a fully connected layer to achieve fine-grained classification of depression states. A loss weighting strategy was employed to improve the model's ability to identify minority classes. Experimental results show that compared with other five-class machine learning models (such as RF, AdaBoost) and two-class combination models (CNN-LSTM, CNN-RNN), CTE-DSR achieves at least 23.0% and 6.7% higher recall and AUC in binary classification, respectively, and at least 15.6% and 1.0% higher F1-score and macro-average AUC in ternary classification. The model has performance advantages in identifying depressive states and can be used to assist in early screening of depression and relapse risk assessment.