A rotational connection between two substructures generally consists of a slewing bearing which is very often used for heavy loads at low speeds. Slewing bearing condition monitoring is a good method to identify the inception and propagation of structural defects at an early stage for timely remedy, and ultimately, enable condition-based "intelligent" maintenance instead of schedule-based. The traditional health evaluation accuracy is not high based on single vibrating signal. We present a new strategy for health evaluation of slewing bearing based on multiple characteristic parameters and the artificial neural networks, and we applied the genetic algorithm to model multi-parameters health evaluation. Seventy days fatigue life test was carried out using a special slewing bearing test table and 70 sets data was chosen to input the networks evaluation model for training and evaluation. Test results show that the ANN (Artificial neural network) with GA (Genetic algorithm) optimization converges more easily, takes fewer iterations and shorter training time, and has higher precision and good stability. And the output of multi-parameters health evaluation network model shows a better agreement with the target. The evaluation results were in accordance with the experiment. Therefore, this methodology can be used to evaluate the health state of slewing bearing.