The BlueSky Smoke Prediction System developed by the U.S. Department of Agriculture, Forest Service, AirFire Team under the National Fire Plan is a modeling framework that integrates tools, knowledge of fuels, moisture, combustion, emissions, plume dynamics, and weather to produce real-time predictions of the cumulative impacts of smoke from wildfires, prescribed fires, and agricultural burn activities. Currently, BlueSky smoke predictions are available daily across the contiguous United States. The output has been used by air regulators, burn bosses, and smoke managers as a guide to help make 'go' and 'no-go' decisions about prescribed fires and plan burn operations. It also helps track day-to-day emissions from wildland and prescribed fires. BlueSky is establishing its reputation as a one-stop shopping for regional smoke concentration and emissions tracking across all land ownership, and is being used by more and more users especially in the West. On the other hand, little is known about the accuracy of its predictions of smoke transport and dispersion under different meteorological conditions. This ongoing study aims at validating BlueSky predictions using in-situ and satellite observations. The study domain is northern California and southern Oregon during the last 2 weeks of August 2006 when several major wildland fires broke out in the region. The predicted smoke concentrations are evaluated by the PM2.5 data at several stations, and the plume trajectories are compared with satellite images. Sensitivity tests are performed to identify potential sources in the smoke predictions so that improvements can be made to the BlueSky prediction system.