Objective To assess the forest health status of Luo Mountain, Ningxia of northwestern China, and to elucidate the driving mechanisms of forest health in arid and semi-arid regions.

Methods This study took five major forest types in Luo Mountain (Picea crassifolia pure forest, Picea crassifolia × Pinus tabuliformis mixed forest, Pinus tabuliformis pure forest, Pinus tabuliformis × Populus davidiana mixed forest, and Populus davidiana pure forest) as the study subjects. By integrating indicators of species diversity, phylogenetic diversity and functional diversity, a forest health evaluation index system was constructed based on four aspects: stand productivity, biodiversity, site conditions and stability. Principal component analysis was employed to calculate the forest health index, and cluster analysis was used to classify forest health levels. Furthermore, regression analysis and elastic network analysis were conducted to investigate the key drivers influencing the health index.

Results (1) Approximately 26.67% of the forests in the study area were in a ‘healthy’ state, 33.33% as ‘sub-healthy’, whilst ‘moderately healthy’ and ‘unhealthy’ each accounted for 20.00%; (2) The health index of mixed forests was significantly higher than that of pure stands (P < 0.05); (3) Standing volume per unit area, species diversity index, phylogenetic diversity index and functional diversity index were the most significant factors influencing forest health; (4) The standardized coefficient values of standing volume per unit area, species diversity index and phylogenetic diversity index were relatively high.

Conclusion Forest health in Luo Mountain is the result of the synergistic interaction between productivity and multidimensional diversity. In the evaluation of forest health in arid and semi-arid regions, the introduction of multidimensional diversity indices helps to enhance the evaluation system’s ability to characterize ecosystem stability and resistance to disturbance.