We present a method to detect drizzle in marine warm clouds by combining visible, near-infrared, and microwave measurements from the Tropical Rainfall Measuring Mission (TRMM) satellite. A visible/near-infrared algorithm is used to simultaneously retrieve cloud optical depth (τ) and effective radius (r _e ). A new microwave algorithm is developed to retrieve liquid water path (LWP _m ). First, the relationship among LWP _m , r _e and τ are investigated by using radiative transfer model simulations and by analyzing satellite observations. Results from both the modeling and the analysis indicate that LWP _m , r _e and τ satisfy a relation in the form of LWP _m = αr _e ^βτ^(3β−1)/2, and that α and β stay relatively constant within a satellite scene for nondrizzle clouds, while they are different from those for drizzle clouds. On the basis of this result, a drizzle detection index (DI) is defined by the ratio of LWP _m to the liquid water path calculated from τ and r _e using a relation derived for nondrizzle clouds. Model simulation results show that for nondrizzle clouds DI is close to 1, and for drizzle/rain clouds DI is greater than 1. Case studies using satellite data indicate that values of DI are greater in the regions with large r _e but relatively small τ, and/or near the center of convective cells. The relation among LWP _m , r _e and τ derived from satellite retrievals is consistent with that of radiative transfer simulation results. Finally, coincident airborne cloud radar measurements were used to verify the capability of DI for drizzle detection. The result indicates that DI is useful in differentiating between drizzle and nondrizzle clouds, and suggesting that DI of ∼1.1 is the drizzle threshold, which is also consistent with our radiative transfer model simulations.