Tamarindus indica is a priority local forest food tree in the Sahel. It is an important source of income for many rural families. It is still in the wild, apart from a few specimens of the accessions recently introduced into the landscape of Sahelian villages. The objective of the present study is to investigate the performance and efficiency of three accession grafts and the effect of grafting to shorten the juvenal growth and fruiting periods for improving woody species diversity in agroforestry parklands and diversifying the sources of income of small producers. The design was a randomized complete block with 5 replications. Each block was composed of 5 grafts of three accessions (Niger 309, TB3, Sweet Thailand) and 5 seedlings propagated from wild tree seeds. Growth parameters (height, collar diameter, canopy width and primary branch number) were measured on all plants. The evaluation of fruiting potential was made on 5 plants of each accession in the first year and 10 plants in the second year. Cost–benefit analysis were carried out to assess the financial viability and economic profitability of planting these accessions of T. indica. The results show that Sweet Thailand accession had the best growth performance compared to Niger 309 and TB3. Fruiting of the three accession grafts started in the second year after planting, in contrast to seedlings which started to bear fruit from the fourth year. TB3 grafts had higher average fruit production in 2017 and 2018 (120 and 640 kg ha−1, respectively). The cost–benefit analysis shows that planting of the three accession grafts is financially viable and economically profitable. The results of the present study indicated that planting these three accessions might play an important role in food security, the resilience of small-scale farmers and the improvement of biodiversity in the landscape.