The invention provides an on-line detection method for blast furnace slag quantity. The method comprises the steps of establishing a fuzzy reasoning model of the blast furnace slag quantity; setting a fuzzy subjection function about (Iti-It0) in combination with the influence characteristic of the magnitude of the (Iti-It0) on the blast furnace slag quantity, wherein Iti is drum current at the ith moment, and It0 is the sum of drum idle current and slag flushing water consumed current; establishing a blast furnace slag quantity calculating model by using the fuzzy reasoning model of the blast furnace slag quantity and the fuzzy subjection function; and performing on-line detection on the blast furnace slag quantity by using the blast furnace slag quantity calculating model. By using the using the fuzzy modeling method, a nonlinear dynamic relation among the water slag quantity, the drum idle current and the slag flushing water consumed current under the condition that the working condition changes is established, and the fuzzy reasoning model of the blast furnace slag quantity is acquired, so that the high-precision on-line detection requirement on the blast furnace slag quantity under the adjustment of the complex working conditions of blast furnace iron-making production can be met, and a reliable judgment basis is provided for smooth operation of a blast furnace.