Intercropping banana crop with a grass cover controlled by mowing is a possible alternative to chemical weed control. However, before adopting this biological way to control weeds, it is crucial to study the impact of the cover crop on banana productivity. In a field experiment, we studied the impact on the banana (Musa spp., cv Cavendish) growth and yield of two grass covers, Cynodon dactylon (Cyn) and Brachiaria decumbens (Bra), managed by mowing every two months, in comparison to conventional weed chemical control (control). The experiment was conducted during two banana cycles. Two months after banana plantating, Bra produced more biomass than Cyn (0.94 kg DM m-2 vs 0.53 kg DM m-2); afterwards both biomass productions became equivalent (0.24 kg DM m-2) when the banana plants shaded the grass. Results showed that these cover crops have impeded the banana growth and development during the first cycle: in comparison to the control, the flowering date was delayed in both cover treatments and the number of fruits per bunch was reduced in the Cyn cover treatment. During the second cycle, as a consequence of the delayed flowering date at the first cycle, banana development was also delayed on Cyn and Bra treatments. However, the presence of Cyn and Bra cover crops had no influence on the size and weight of the banana plants at flowering date, nor on the final crop productivity. These results were interpreted as a result of contrasted conditions of competition for nutrient during the two cycles: an important production of biomass and nutrient demand of the two cover crops during the first cycle, impacting the banana growth and delaying the flowering date and, during the second cycle, a lower biomass production of the both covers crops shaded by the banana canopy and a contribution to the nutrient supply provided by the decomposition of the mowed residues.