Abstract: Aiming at the joint optimization problem of location layout and capacity allocation of pick-up/drop-off points (P/D point) in cellar flow shops, an optimization model with throughput rates and cycle time constraints is established with the objective of minimizing the total transportation cost (including congestion cost) and allocation cost. According to the capacity balance between P/D points of adjacent processes and the problem characteristics of joint optimization of P/D point locations and capacities, a mesh adaptive direct search algorithm embedded with adaptive cooperative neighborhood search algorithm (ACNS-MADS) is proposed, in which the ACNS is used to reoptimize the P/D point location layout and capacity allocation scheme of a new solution. Experimental results show that compared with other algorithms, the total transportation cost and the P/D point allocation cost obtained by ACNS-MADS are reduced by 2.99% and 5.64% respectively, with a computation time reduction of over 17.95%. It can conclude that the proposed algorithm is effective and efficient to solve the joint optimization problem of P/D point layout and allocation, having practical value.