This paper analyses the process of adoption of no tillage in South-eastern Spain’s olive groves. Olive tree groves in South-eastern Spain’s mountainous areas are subject to a high risk of soil erosion, which is the main environmental problem for this crop, and have to incur in high costs of soil conservation. This results in a greater difficulty to comply with the practices required to benefit from both the single payment and agri-environmental schemes. In many high-steeped areas, farmers have opted for non-tillage practices as an alternative to other conservation practices. Using our own data from a survey carried out in 2006 among 215 olive tree farmers from the Granada Province in Southern Spain regarding the adoption of soil conservation and management practices, we model the diffusion process of no tillage practices using several specifications (logistic, Gompertz and exponential). We also estimate an ordered probit model to analyse which socio-economic and institutional factors determine the adoption of no tillage. Our results show that 90% of farmers in the area of study perform no tillage with either localized (21%) or no localized (69%) application of weedicides. The diffusion process of no tillage has been very intense since the middle nineties, and has been based on the interactions among farmers in the area of study rather than in external factors such as EU policies or extension services. Among other relevant factors that positively affect the adoption of no tillage practices in general, such as farm size and irrigation, the probability of a farmer adopting no tillage with non-localized application of weedicides increases when there is a relative that will continue with the farming activity, what causes the farmer to incorporate long term effects in his farming decisions, when the farmer is only a manager or when he bought the farm rather than inherited it (i.e. on more professionalized farms), and with his educational level. These results confirm some findings from previous studies in other nearby areas.