The measurements of net particulate primary production (NPP) with reference to space and time are limited in the Indian Ocean. Therefore, models are the only source of understanding spatial, temporal, and long-term variations. Different models have been developed to use remote sensing data to estimate NPP in the Indian Ocean. However, the accuracy of these models has not been evaluated. Here, we used the measured NPP data collected between 2003 and 2022 in the Indian Ocean to compare with different models such as Vertically Generalized Production (VGPM), Carbon-based Production (CbPM), Eppley-VGPM, and Carbon, Absorption, Fluorescence, Euphotic resolved (CAFE) models. This comparison suggested that CbPM simulated NPP with 10% underestimation compared to measured NPP in the Arabian Sea and South Indian Ocean (SIO), whereas VGPM simulated NPP well in the Bay of Bengal (BoB). The Eppley-VGPM over (under) estimated measured NPP by 20–30% in the SIO (Arabian Sea), while the CAFE model underestimated NPP in the BoB by 25%. No single model simulated the NPP well in the entire Indian Ocean compared to the measurements. All models reproduced seasonality in the Arabian Sea and SIO, but they failed in the case of the BoB. Compared to the measured NPP, the VGPM and CAFE models underestimated NPP, whereas the CbPM and Eppley-VGPM models overestimated it. Variable rates of changes in NPP in the past two decades were simulated in the Indian Ocean by different models. The long-term trends simulated by CbPM suggest decreasing trends in NPP south of 12°N and no trends north of 12°N in the Arabian Sea and increasing trends in NPP south of 20°S and decreasing trends north of 20°S in the SIO. The VGPM model indicated no significant trends in the BoB. The trends by the other models are variable and contrast in some regions. Therefore, the selection of an appropriate model to study long-term trends in NPP in the Indian Ocean is crucial.