This paper presents a mathematical model for task allocation optimization problems, considering parallel workstations and task paralleling. A mixed integer programming is formulated to take into account operational constraints to minimize the cycle time. The model is then applied to solve task allocation for assembly line balancing in the textile and garment industry, including ten operations and four types of machines. The results show a significant increase in line efficiency, from 52.47% to 94.20%, and a 45% reduction in cycle time. This research contributes to exploring the capabilities of optimization models for line balancing under parallel workstations and task-paralleling constraints that a limited number of researchers have not yet thoroughly investigated.