This paper investigates tracking control in the rotational angle domain based on the time-varying internal model principle. The objective is to enable precise, reliable, and computationally efficient output tracking of signals that are dependent on angular displacement. To achieve desired performance, existing approaches based on internal model principle require a large number of samples per revolution, which significantly increases the controller order and also poses challenges for the transient performance. To address those issues, a varying sampling interval approach is proposed, where the angular sampling locations are not fixed but optimized based on tracking errors between sampling points so that desired performance can be achieved without increasing the number of samples. Meanwhile, to improve the convergence rate of the tracking error, additional linear matrix inequalities (LMI) constraints are added to the existing stabilizer synthesis. Through experimental study on a camless engine valve actuation system, the effectiveness of the proposed approaches is demonstrated. It is shown that, compared with the fixed interval sampling, the varying sampling approach can reduce the tracking error by over 50%.