This paper proposes a Multi-Objective Teaching-Learning-Based Optimizer (MOTLBO) algorithm to solve the Multi-Weeding Robot Task Assignment (MWRTA) problem, which aims to minimize the maximum completion time and residual herbicide in agricultural robots. The MWRTA problem is modeled as a multi-robot task assignment problem, where multiple agricultural robots are assigned to perform weeding tasks in a smart farmland. The algorithm uses a mathematical model to define the problem and introduces a multi-objective optimization approach to find optimal or near-optimal solutions. The MOTLBO algorithm is designed with a heuristic-based initialization using an improved NEH heuristic and a maximum load-based heuristic to generate a high-quality and diverse initial population. A dynamic grouping mechanism and a redefined individual updating rule are introduced to enhance global exploitation and prevent local optima. A multi-neighborhood-based local search strategy is also provided to balance the algorithm's exploration and exploitation. The algorithm is compared with several state-of-the-art algorithms, and the results show that the proposed algorithm outperforms them in terms of performance metrics such as Inverse Generation Distance (IGD) and Hypervolume (HV). The algorithm is effective in solving the MWRTA problem, which is crucial for improving the efficiency and cost-effectiveness of smart farming. The key contributions of this paper include the development of a mathematical model for the MWRTA problem, the design of the MOTLBO algorithm, and the introduction of a multi-neighborhood-based local search strategy. The algorithm is validated through extensive experiments, and the results demonstrate its superiority in solving the MWRTA problem.This paper proposes a Multi-Objective Teaching-Learning-Based Optimizer (MOTLBO) algorithm to solve the Multi-Weeding Robot Task Assignment (MWRTA) problem, which aims to minimize the maximum completion time and residual herbicide in agricultural robots. The MWRTA problem is modeled as a multi-robot task assignment problem, where multiple agricultural robots are assigned to perform weeding tasks in a smart farmland. The algorithm uses a mathematical model to define the problem and introduces a multi-objective optimization approach to find optimal or near-optimal solutions. The MOTLBO algorithm is designed with a heuristic-based initialization using an improved NEH heuristic and a maximum load-based heuristic to generate a high-quality and diverse initial population. A dynamic grouping mechanism and a redefined individual updating rule are introduced to enhance global exploitation and prevent local optima. A multi-neighborhood-based local search strategy is also provided to balance the algorithm's exploration and exploitation. The algorithm is compared with several state-of-the-art algorithms, and the results show that the proposed algorithm outperforms them in terms of performance metrics such as Inverse Generation Distance (IGD) and Hypervolume (HV). The algorithm is effective in solving the MWRTA problem, which is crucial for improving the efficiency and cost-effectiveness of smart farming. The key contributions of this paper include the development of a mathematical model for the MWRTA problem, the design of the MOTLBO algorithm, and the introduction of a multi-neighborhood-based local search strategy. The algorithm is validated through extensive experiments, and the results demonstrate its superiority in solving the MWRTA problem.