Electroencephalogram (EEG)-based emotion decoding can objectively quantify people’s emotional state and has broad application prospects in human-computer interaction and early detection of emotional disorders. Recently emerging deep learning architectures have significantly improved the performance of EEG emotion decoding. However, existing methods still fall short of fully capturing the complex spatiotemporal dynamics of neural signals, which are crucial for representing emotion processing. This study proposes a Dynamic-Attention-based EEG State Transition (DAEST) modeling method to characterize EEG spatiotemporal dynamics. The model extracts spatiotemporal components of EEG that represent multiple parallel neural processes and estimates dynamic attention weights on these components to capture transitions in brain states. The model is optimized within a contrastive learning framework for cross-subject emotion recognition. The proposed method achieved state-of-the-art performance on three publicly available datasets: FACED, SEED, and SEED-V. It achieved <inline-formula><tex-math notation="LaTeX">$81.7\pm 4.3\%$</tex-math><alternatives><mml:math><mml:mrow><mml:mn>81</mml:mn><mml:mo>.</mml:mo><mml:mn>7</mml:mn><mml:mo>±</mml:mo><mml:mn>4</mml:mn><mml:mo>.</mml:mo><mml:mn>3</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="zhang-ieq1-3593630.gif"/></alternatives></inline-formula> accuracy in the binary classification of positive and negative emotions and <inline-formula><tex-math notation="LaTeX">$67.9\pm 7.3\%$</tex-math><alternatives><mml:math><mml:mrow><mml:mn>67</mml:mn><mml:mo>.</mml:mo><mml:mn>9</mml:mn><mml:mo>±</mml:mo><mml:mn>7</mml:mn><mml:mo>.</mml:mo><mml:mn>3</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="zhang-ieq2-3593630.gif"/></alternatives></inline-formula> in nine-class discrete emotion classification on the FACED dataset, <inline-formula><tex-math notation="LaTeX">$88.1\pm 3.6\%$</tex-math><alternatives><mml:math><mml:mrow><mml:mn>88</mml:mn><mml:mo>.</mml:mo><mml:mn>1</mml:mn><mml:mo>±</mml:mo><mml:mn>3</mml:mn><mml:mo>.</mml:mo><mml:mn>6</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="zhang-ieq3-3593630.gif"/></alternatives></inline-formula> in the three-class classification of positive, negative, and neutral emotions on the SEED dataset, and <inline-formula><tex-math notation="LaTeX">$73.6\pm 12.7\%$</tex-math><alternatives><mml:math><mml:mrow><mml:mn>73</mml:mn><mml:mo>.</mml:mo><mml:mn>6</mml:mn><mml:mo>±</mml:mo><mml:mn>12</mml:mn><mml:mo>.</mml:mo><mml:mn>7</mml:mn><mml:mo>%</mml:mo></mml:mrow></mml:math><inline-graphic xlink:href="zhang-ieq4-3593630.gif"/></alternatives></inline-formula> in five-class discrete emotion classification on the SEED-V dataset. The learned EEG spatiotemporal patterns and dynamic transition properties offer valuable insights into neural dynamics underlying emotion processing.