To solve multi-camera pose estimation problem with the coexistence of point and line segment under the industry visual scene, a new pose estimation algorithm for multi-camera from point and line correspondences is proposed. The method uses geometric restriction between multiple calibrated cameras, and globally processes point and line segment feature from overlapping and non-overlapping views of multi-camera rig, finally obtains more robust and accurate pose estimation. In the proposed method, the formulation of datum of multiple cameras is unfined, and the sum of object-space collinearity error of point and coplanarity error of line segment all cameras are used as error function. Finally, an iterative algorithm is developed to minimize the error function. The experimental results prove the effectiveness and superiority of the algorithm for pose estimation.