Collaborative optimization is an important and active multidisciplinary design method in the current complex vehicle system engineering. However, there are many problems when the traditional collaborative optimization is applied in vehicle body weight reduction, such as convergence difficulties and low optimization efficiency. To solve these problems, an improved multidisciplinary design and collaborative optimization algorithm is proposed in the paper. Firstly, an adaptive relaxation factor and an adjustable penalty are introduced to convert the system level constraint optimization problem into unconstraint optimization problems. The improved algorithm can overcome the disadvantages of convergence difficulties of traditional collaborative optimization. Then, a classical example is used to verify the effectiveness and stability of the proposed algorithm. Finally, the improved multidisciplinary design and collaborative optimization algorithm is applied in vehicle body weight reduction. The architecture of body lightweight design considering safety performance and NVH performance using the proposed algorithm is established and run. As is shown in results, the weight of vehicle body with optimization design is reduced under the premise of satisfying requirements of safety performance and NVH performance.