<p>Line balancing involves assigning tasks to workstations in a production line. This study focuses on the robotic assembly line balancing problem type-II (RALBP-II), which aims to minimize cycle time with a fixed number of workstations while assigning robotic arms to workstations. Because cycle time evaluation must be repeatedly performed during the search process, efficient cycle time calculation and effective optimization strategies are critical for solving this problem. To address this issue, this study first proposes an improved cycle time method (ICTM), which dynamically adjusts the cycle time search range to improve calculation efficiency. In addition, a hybrid particle swarm optimization (HPSO) algorithm is developed by integrating particle swarm optimization with order crossover (OX) and block mutation (BM) to enhance local search capability and improve solution quality. For the comparison of cycle time calculation methods, the largest benchmark instance with 297 tasks is adopted as the test case. The results show that ICTM outperforms the compared cycle time calculation methods. Furthermore, the proposed HPSO algorithm is evaluated on eight benchmark instances ranging from 25 to 297 tasks. The computational results indicate that HPSO achieves better solution quality than genetic algorithms, particle swarm optimization, ant colony optimization, PSO_OX, and PSO_BM.</p>