The present study aims to investigate the effect of foreign direct investment on total factor productivity by controlling for institutional factors, comparing two groups of developed and developing countries using panel data of 48 countries, including 29 developed countries and 19 developing countries, during the period 2000 to 2024. The model estimation was performed using the threshold panel regression (PTR) approach. The results of the model estimation showed that the nature of the effect of FDI on TFP in the two groups of countries is completely different. In developed countries, the relationship between FDI and productivity is linear and the FDI coefficient is statistically insignificant, which is explained by the technological frontier hypothesis and the reliance of these countries on endogenous innovation. The results for developing countries indicate the existence of a strong nonlinear and threshold relationship; so that the research and development variable divides the sample into three separate regimes by creating two threshold values ​​of 0.062 and 0.233 percent of GDP. The findings show that, although the absolute dependence on foreign technology decreases after passing the R&D thresholds, the positive and complementary effect of FDI on productivity continues. The effect of FDI on TFP in the first regime (0.093) is much stronger than in the second regime (0.061) and the third regime (0.046), which is fully consistent with the absorptive capacity theory. Also, the quality of governance has the strongest positive effect in developing countries with a coefficient of 0.228, while it is insignificant in developed countries. Also, the variables of human capital and financial development had an asymmetric effect on productivity in the two groups of countries.