The Himalayan region is adversely affected by the increasing anthropogenic emissions from the adjacent Indo-Gangetic plain. However, source apportionment studies for the Himalayan region that are crucial for estimating CO concentration, are grossly insufficient, to say the least. It is in this context that our study reported here assumes significance. This study utilizes five years (2014–2018) of ground-based observations of eBC and multiple linear regression framework (MLR) to estimate CO and segregate its fossil fuel and biomass emission fractions at a high-altitude (1958 m) site in the Central Himalayas. The results show that MERRA2 always underestimates the observed CO; MOPITT has a high monthly difference ranging from −32% to +57% while WRF-Chem simulations underestimate CO from February to June and overestimate in other months. In contrast, CO estimated from MLR replicates diurnal and monthly variations and estimates CO with an r2 > 0.8 for 2014–2017. The CO predicted during 2018 closely follows the observed variations, and its mixing ratios lie within ±17% of the observed CO. The results reveal a unimodal diurnal variation of CO, COff (ff: fossil fuel) and CObb (bb: biomass burning) governed by the boundary layer evolution and upslope winds. COff has a higher diurnal amplitude (39.1–67.8 ppb) than CObb (5.7–33.5 ppb). Overall, COff is the major contributor (27%) in CO after its background fraction (58%). CObb fraction reaches a maximum (28%) during spring, a period of increased agricultural and forest fires in Northern India. In comparison, WRF-Chem tracer runs underestimate CObb (−38% to −98%) while they overestimate the anthropogenic CO during monsoon. This study thus attempts to address the lack of continuous CO monitoring and the need to segregate its fossil fuel and biomass sources, specifically over the Central Himalayas, by employing a methodology that utilizes the existing network of eBC observations.