Conservation agriculture (CA) is rapidly developing in Madagascar but little is known about its effects on local soil functioning. To assess some of those effects, we investigated the effects of three CA systems and two levels of fertilization on soil functioning using nematofauna as indicator. The systems consisted in (i) soybean (Glycine max L.)-maize (Zea mays L.) rotation with mulch of residues, CA-R; (ii) bean (Phaseolus vulgaris L.)-soybean rotation with living mulch of Pennisetum clandestinum, CA-K; (iii) continuous maize with living mulch of Desmodium uncinatum, CA-D and were compared with soybean-maize under conventional tillage (CT) and natural fallow (NF). The fertilization levels consisted in (i) farmyard manure, FYM; and (ii) farmyard manure + mineral fertilizers, FYM + NPK. Located in the Highlands of Madagascar, the experiment was setup in 1991 and andic Dystrustept soil had been sampled in 2005-2007. We measured nematode abundances and ecological indices as well as the abundance and biomass of soil macrofauna, soil water and organic C and N contents and plant yields. We hypothesized that (1) CA including maize in monoculture would lead to higher abundance of plant-parasitic nematodes; (2) both dead-residue mulch and inorganic fertilization would lead to a more basal nematode community structure; and (3) that the combination of system effects on soil nematode community would be able to forecast differential crop yields for the CA systems. Our results show that CA systems tested were able to support better/comparable maize and soybean yields compared with CT, provided that crop rotation is correctly managed. Supporting our first hypothesis, abundance of plant parasitic nematodes was (40-150 times) higher under maize monoculture. Abundance of soil nematofauna and trophic groups (excepted carnivores and omnivores) was stable during the three years. Inorganic fertilization increases carnivorous and omnivorous nematodes to 122% and 140%, respectively. Ecol