The simple monkey algorithm has the shortcomings of initial distribution randomization, fixed climb step length and incapability of inheriting the characteristics of excellent monkeys, which limits the solution performance of the algorithm. In order to solve the above problems, a weed monkey algorithm for optimal sensor placement is proposed. Normal distribution is used to enhance the diversity of initial monkey populations. Selfadaptive climb step is introduced to improve the solution accuracy and convergence rate. Both weed reproduction evolution and competitive exclusion mechanism are used to enlarge the influence of excellent monkey on the monkey offspring population. The commonly used Modal Assurance Criterion (MAC) is used as the objective function of optimal sensor placement. The commonly used 8 test functions and 3 algorithms are used to verify the feasibility and effectiveness of the algorithm. Finally, the optimal sensor placement is carried out on the gelatinize mechanism of bag bottompasting machine. The results show that, compared with the simple monkey algorithm, the solution accuracy of the weed monkey algorithm precision is greatly improved.