We present the Adaptive Hybrid Cursor, a novel target acquisition technique for pen-based interfaces. To assist a user in a target selection task, this technique automatically adapts the size of the cursor and/or its contexts (the target size and the selection background) based on pen pressure input. We systematically evaluated the new technique with various 2D target acquisition tasks. The experimental results indicated that the Adaptive Hybrid Cursor had better selection performance, and was particularly effective for small-target and high-density environments in which the regular cursor and the Bubble Cursor [13] failed to show significant advantages. The Adaptive Hybrid Cursor is a novel way to improve target acquisition via pressure input, and our study demonstrated its viability and potential for pen-based interfaces.