A continuous-time mixed integer linear programming model and an improved two-dimensional particle warm optimization(PSO) algorithm are designed to tackle integrated lotsizing and scheduling for multi-variety batch production by turns in chemical industry after analyzing the characteristics of multistage, shared equipments, material input and output fluctuant conversion rate, inventory limitation and product changeover. The coding scheme of particles is designed in terms of the processing state of production units, while an effective decoding procedure translates a particle into a feasible lotsizing and scheduling solution. The improved PSO algorithm incorporates contraction operators to improve the intensification ability of the algorithm. In addition, divergence operators and velocity disturbance strategies are also introduced into the PSO algorithm to keep the diversity of the swarm. Computational results show the good performance of the proposed PSO algorithm.