Radiotherapy sessions often last several minutes, during which patients continue to breathe, leading to significant intrafraction organ motion. Accurately predicting and modeling this motion is crucial for precise dose delivery. However, existing pre-treatment organ motion prediction methods primarily rely on deformation analysis using principal component analysis (PCA), which is highly dependent on registration quality and struggles to capture periodic temporal dynamics for motion modeling.In this paper, we observe that organ motion prediction closely resembles an autoregressive process, a technique widely used in natural language processing (NLP). Autoregressive models predict the next token based on previous inputs, naturally aligning with our objective of forecasting future organ motion phases. Building on this insight, we reformulate organ motion prediction as an autoregressive task to better capture patient-specific breathing dynamics. Specifically, we acquire 4D CT scans for each patient before treatment, with each sequence comprising multiple 3D CT phases. These phases are fed into an autoregressive model to generate future motion phases, directly predicting images without requiring ground-truth deformation vector fields (DVFs). We evaluate our method on a real-world test set of 4D CT scans from 50 patients who underwent radiotherapy at our institution and a public dataset containing 4D CT scans from 20 patients, totaling over 1,300 3D CT phases. Our model outperforms existing benchmarks in predicting lung and heart motion, demonstrating its ability to implicitly capture clinically meaningful motion dynamics from image sequences. These results highlight the potential of our approach to enhance pre-treatment planning in radiotherapy, particularly in scenarios where DVFs are unavailable or unreliable.