Electroencephalogram (EEG) contains a wealth of temporal, spatial and frequency information, which is the most accurate physiological signal to reflect the emotional state, and plays an important role in the field of emotion recognition.Because the single feature EEG emotion recognition method has the problem of missing information, this paper proposes a three-dimensional fusion feature EEG information processing method, which fuses the differential entropy frequency domain features and eight kinds of time domain features of EEG signals, arranges the EEG signals according to the electrode position information, and extracts the three-dimensional mixed features of EEG signal.The attention mechanism is introduced into the multi-task convolutional neural network (MTA-CNN), and the constructed three-dimensional features are used as the input for test and analysis. The results show that the proposed model and method can significantly improve the classification accuracy in the valence dimension and arousal dimension of DEAP data set.