In this paper, we present new Proxy Means Test (PMT) models for targeting based on the 2017 household survey data (EDAM4). The paper finds that the developed PMT models, with separate targeting formulas for rural and urban areas, appear to perform well vis-a-vis inclusion and exclusion errors observed in similar country contexts. Ex-ante simulations also show that the planned expansion of targeted social assistance program, PNSF, using the proposed targeting approaches will result in nearly 0.7 percentage point reduction in poverty nationally and 3.4- 4.1 percentage reductions in the regions. These results reinforce the effectiveness of PMT targeting approach for social assistance programs in Djibouti when carried out in combination with geographic targeting in high-poverty districts. The results also show the relevance and effectiveness of PMT as a national targeting approach in urban regions and for an expanded national social assistance program.