With the continuous shortening of the vehicle development cycle

the traditional vehicle structure design methods are difficult to meet the rapid iteration of the market. Finite element simulation analysis can accurately evaluate the structural performance of vehicles

but it consumes a lot of computing time and labor costs. Aiming at the above problems

this paper takes a certain type of automobile engine hood as an example

selects sample points in the design space through DOE design of experiments

and constructs a co-simulation analysis process using Isight and Hypermesh

which shortens the manual simulation cycle. Based on the co-simulation data

the Stacking ensemble learning model is used to construct the performance prediction model of the car hood

which realizes the prediction of mass

torsional stiffness and first-order mode

and verifies the effectiveness of the prediction model. The research result shows that the prediction model constructed by machine learning method can accurately evaluate the performance of automobile hood

which provides a new path for the intelligent design of automobile structure.