Aiming at the impact of bearing bushes on the lubrication and friction and wear of diesel engine connecting rod big-end bearings, a certain type of diesel engine connecting rod big-end bearing bush was taken as the research object and a multi-body dynamics numerical calculation model of the connecting rod group was built. The connecting rod big-end bearing bushes with four profiles: exponential, hyperelliptic, barrel and trapezoid were used to study the effect on bearing lubrication. The study found that the hyperelliptic bush has the best lubrication performance for the connecting rod big-end bearing. On the basis of the hyperelliptic bush, the bearing clearance, bearing width, journal oil hole diameter and oil supply pressure are used as design variables, using the Box–Behnken experimental design and radial basis function (RBF) neural network method to construct an approximate multi-objective model, which the minimum oil film thickness (abbreviated as MOFT) and average rough contact effective pressure are the objectives. A non-dominated sorting genetic algorithm (NSGA-II) is used for multi-objective optimization. The optimization results show that the bearing clearance remains basically unchanged, the bearing width, journal oil hole diameter and oil supply pressure increase, so that MOFT of the connecting rod big-end bearing rises from 1.56 μm to 1.97 μm, and the average rough contact effective pressure increases from 3.97 MPa decreases to 0.25 MPa. The research results can provide a reference for the analysis and optimization of the lubrication performance of the connecting rod big-end bearing.