A real-time prediction method for an engine emission is provided, including: acquiring multiple known historical test data samples for engine emission, dividing the samples into a training set and a test set to train multiple neural network (NN), calculating mean square errors (MSEs) output by each of the NNs with the different numbers of hidden layer nodes to determine a topological structure of the NNs, optimizing initial weights and initial thresholds for each of the NNs with a mind evolutionary algorithm (MEA), and establishing a real-time engine emission prediction system with an Adaboost algorithm. The present disclosure overcomes problems that the existing engine emission data acquisition method is time-consuming and labor-consuming, restricted by environmental factors, expensive in instrument cost and undesirable in transient emission measurement performance; and only by simply measuring a rotational speed a torque, a power, a rail pressure, an air-fuel ratio, an oil consumption, an exhaust gas recirculation (EGR) rate and a start of injection (SOI) in operation of the engine, the present disclosure can accurately predict the transient nitrogen oxide (NOx) emission, total hydro carbon (THC) emission and carbon monoxide (CO) emission of the engine in real time.