Vineyard soils host diverse arthropod communities, including Mesostigmata mites, that are abundant but understudied soil predators in agroecosystems. The effects of viticultural practices, particularly plant diversification, on these predator communities remain poorly documented despite their importance for pest regulation. By modifying habitat structure, vegetation cover, and resource availability, plant diversification may shape soil predator communities and influence their ecological functions. In this context, the following questions were addressed: (i) How do management practices related to plant diversification affect the abundance of soil arthropod groups especially Mesostigmata? (ii) How do sampling period and spatial position within rows shape Mesostigmata communities? (iii) Does plant diversification affect Mesostigmata through bottom-up effects by increasing vegetation cover diversity, soil stability, and prey availability? We compared mesofauna communities, focusing on Mesostigmata, along a vegetation diversification gradient in an experimental Mediterranean vineyard (Domaine du Chapitre, Villeneuve-lès-Maguelone). This gradient included conventionally tilled plots (TViti), cover-cropped plots (AViti), and agroforestry plots with spontaneous cover crops and fruit trees replacing central vine rows (DViti). Grape variety is Artaban (resistant to fungal diseases). In total, 324 soil samples were collected in April, June and July 2024 from both rows and inter-rows. Microarthropods were extracted using Berlese–Tüllgren funnels. Mesostigmata were identified to family and species levels. Data were analysed using GLMMs, ordination methods and PLS-PM. Of the 55,983 arthropods sorted, 71% were mites. We counted 1,957 Mesostigmata, including over 25 morphotypes from ten families. Cover vegetation increased the abundance of several arthropod groups, including Mesostigmata, Oribatida, Trombidiformes, Collembola and ants, with generally higher abundances in AViti and DViti than in TViti. Mesostigmata abundance varied with sampling period and position and tended to weaken later in the season, with the highest abundance generally observed in the inter-row. Mesostigmata species richness was higher under cover vegetation (AViti and DViti) than under bare soil (TViti), decreased over the sampling period, and was higher in the inter-row than in the vine row. Community composition showed rather weak structuring, with a small effect of position and a management × sampling period interaction in NMDS, whereas constrained ordination (dbRDA) explained little variation (R²adj = 0.05) and revealed no clear separation among treatments. PLS-PM for the second sampling season revealed a direct effect of vegetation diversification on plant community diversity. The total effects indicated that vegetation diversification seems to positively influence soil stability, prey communities and mesostigmatid mites.