Soil texture is well known to directly affect bioavailability of organic matter to heterotrophs, but it also steers their activity by moderating soil moisture fluctuation. Disentangling these direct and indirect textural controls is, however, not trivial and attempts to do so are very scarce. Most attention has just gone to the stimulation of soil carbon (C) mineralization by soil moisture fluctuation per se. To quantify the indirect moisture-mediation control of soil texture on C mineralization, we monitored maize straw degradation in various soil texture/moisture regime combinations. Moisture levels were firstly kept fixed at 32% WFPS (experiment Fixed32) in a sand, sandy loam and silt loam soil or allowed to fluctuate between 20% and 50% water-filled pore space (WFPS, Dry-wet2050). Total maize-C (Cmaize) mineralized was highly similar between these three textures and thus the direct textural control was minor. On the contrary with the fluctuating moisture level, around threefold more added Cmaize was mineralized (P < 0.01) in the sand (86%) than in the silt loam (25%) soil. We owe this boost in Cmaize mineralization to the rewetting of larger pores in the sandy soil which should contain most of the ground maize residue. This determining control of texture on distribution of moisture as well as the maize substrate clearly exceeded the direct impact of texture on organic matter stability. In a third and fourth experiment, timing and dose to remoisten a silt loam soil back to 50% (experiment Equal20-50) or 35% WFPS (experiment Equal20-35) when dried out to 20% WFPS were additionally mimicked in the sand and sandy loam soils. These scenarios correspond more closely to a field situation in which nearby differently textured soils all receive the same precipitation input. After several rewetting cycles, the silt loam soil eventually had a 13.9% higher WFPS than the sandy soils. By then moisture stress clearly limited Cmaize mineralization in the sandy soil (0.13 (Equal20-50) and 0.05 (Equal20-35) mg kg−1 h−1) as it proceeded at only half of the rate as in the silt loam soil (0.21 (Equal2050) and 0.12 (Equal20-35) mg kg−1 h−1). The amount of Cmaize mineralized after 120 days in the silt loam soil was simulated to surpass that in the sandy soil. We conclude that the effect of soil texture on decomposition of a fresh substrate is largely indirect, i.e. through mediation of soil water content and its distribution in the soil matrix. Moreover, our data suggests that in the event of prolonged drought Cmaize mineralization will be less limited in finer textured soil, contradicting the widespread idea that organic matter would degrade more rapidly in coarser textured soils.