The purpose of this research is to evaluate the results of a multisensory/phonic literacy intervention conducted with Brazilian children in last stage of kindergarten who present risk for learning difficulties in reading and writing. Participants are 36 children, with average age of five, students of a public school. The experimental design consisted of the following stages: pre-test, intervention, post-test one and post-test two. The research is ongoing, therefore, the partial data presented correspond to pre-test and post-test one stages. In pre-testing the children were assessed through evidence of letter knowledge, phonological awareness, reading and writing surveys. Children who had lower results than the average of the group in letter knowledge and phonological awareness and presented no understanding of the alphabetic principle in their attempts at reading and writing, were considered at risk (n=15). The children at and out of risk were divided randomly into two groups: Multisensory Group (MG) and Control Group (CG). Each group had two sub-groups formed by children at risk (MGr and CGr) and children out of risk (MGo and CGo). It was found no significant performance differences between the groups MG and CG. The Multisensory Group participated in three weekly intervention sessions lasting 35 minutes, during 16 weeks. The program consisted of exercises stimulating phonological awareness and regular correspondence between letter and sound through the use of multisensory stimuli (auditory, visual, kinesthetic and tactile), including motion with the whole body to represent the letter's shape. The Control Group sessions were designed by placebo activities (storytelling and recreational activities) and occurred with the same length of Multisensory Group sessions. Post-test one was performed after four months. Statistical analyzes assessed the significance results obtained with intervention. Comparisons between pre-test and post-test one showed a significant difference for all abilities evaluated in MGr : letter knowledge, phonological awareness, reading and writing. The MGo presented significant difference for letter knowledge and phonological awareness. Furthermore, the performance of CGr and CGo presented significant difference only in letter knowledge. The data also revealed a significantly higher performance of MGr for all abilities assessed. Moreover, the performance of MGo wasn't significantly different from CGo for any abilities evaluated. These results show an important effect of multisensory method with phonic emphasis to improve important abilities involved in literacy in children at risk for presenting learning difficulties in reading and writing. However, the question about the stability of these results still remains.