The diversity index is often calculated based on the data of species identification level. If the diversity index at the species level can be replaced by high-level taxonomic identification results (e.g., families) or directly by the number of individuals within the indicator group, it can greatly save the cost and improve the work efficiency. Field sampling data from multiple indicator taxa and multiple analytical methods could be used to verify this problem, which is helpful to obtain more general conclusion. In order to study the substitution effect of different classification levels and individuals of arthropods on species richness, arthropod sampling was carried out in two different management measures and multi-habitat farmland landscape areas in Ningbo, Zhejiang Province from May to August, 2019. Arthropods were identified to order level, while spiders and bees were identified at both family level and species level. The correlations between the number of taxa at order or family level, the number of individuals in indicator taxa and species richness were analyzed, and two-way analysis of variance (ANOVA) based on these four indexes between different management measures and different habitats were carried out. The non-metric multidimensional scale analysis (NMDS) was used to compare their species composition differences between different management measures and habitats based on the data of order, family and species level. The results showed that:(1) 63% and 89% of the variability of species richness based on species level data could be fitted by the number of family of spiders and bees, respectively, and the correlation between species richness and the number of family of spiders or bees (r=0.79 and r=0.95, respectively) was greater than that of arthropod order level data (r=0.32 and r=0.53, respectively). (2) There was a very significant correlation (r ≥ 0.7) between the number of individuals and species richness within the indicator taxon. (3) The two-way ANOVA showed that there were significant differences between different management measures based on three identified level data, but the significant differences among habitats based on the data of family and individuals were more similar to those of species richness than the data of order level. (4) The results of NMDS based on data of family level were also more similar to those of species level than the order level. There were significant differences of species or family composition of spiders or bees among different management measures or some habitats, but not in arthropod order composition. Since the higher classification levels such as the order, are not natural units, but artificial classification levels facilitate the classification of information systems, the results may not better reflect the indicator results of species richness. Therefore, for the assessment of farmland biodiversity, the classification data at a higher level could be used to some extent, especially the family level would be the first choice. If a rapid assessment of farmland biodiversity based on a large number of spider or bee specimens is required urgently, the number of individuals of indicator taxon can be directly counted to replace the results of species richness.