为了保证海啸预警的时效性，复杂的地震震源经常被简化为均匀滑移模型来预测海啸波. 虽然均匀滑移模型已经被广泛使用，但其在实际事件中预测海啸波的准确性并未得到全面的评估和认可.对2021年M _w 8.2 Alaska地震构建了有限断层模型（finite-fault model）和多种均匀滑移模型，并对海啸波的预测误差进行对比分析.有限断层模型显示，2021年Alaska地震的同震滑移分布在15~40 km的深度范围内，震源周围的最大滑移约为6 m. 另外，通过全局搜索得到的最优均匀滑移模型对海啸波的预测与有限断层模型非常接近，都与观测波形符合良好；两种位于gCMT中心、但采用不同标度关系（scalingrelation）的均匀滑移模型给出了几乎一致的远场波形.对此次地震海啸的研究结果表明，均匀滑移模型对海啸波的最佳预测能力与有限断层模型相当，根据gCMT中心和标度关系构造的均匀滑移模型对远场海啸预警比较可靠，且不同标度关系对远场波形预测无显著影响.

To issue tsunami warnings in real-time, complex earthquake sources are usually simplified to uniform slip models for tsunami prediction. Althought this approach of simplification is widely used, its accuracy in predicting tsunami waves in actual events hasnot been fully evaluated and recognized. In this paper, a finite-fault model and various uniform slip models are constructed for the 2021 M _w 8.2 Alaska Peninsula earthquake, and their prediction errors for tsunami waves are compared. The finite-fault model inverted from tsunami datareveals that the coseismic slip of this event was distributed over a depth range of 15 to 40 km, and the ~6m maximum slip occurred near the hypocenter. Besides, the optimum uniform slip model obtained from global search provides very similar tsunami predictions to those given by the finite-fault model, both of which agree well with the observations. Two uniform slip models located at the gCMT centroid but using different scaling relations yield almost the same far-field waveforms.Results of this study show that the optimum predicting ability of uniform slip modelis almost equivalent to that of the finite-fault model. The uniform slip models based on gCMT centroids and scaling relations are relatively reliable for far-field tsunami warning, and difference in scaling relations may not significantly impact the far-field predictions.