Owing to the lack of direct measurement on the slagging extent of the waterwall, random or empirical soot-blowing strategies practiced in many power plants can result in untimely or excessive soot-blowing operations. In this research, a dynamic slagging monitoring model was established based on the heat balance principle and GA-BP (genetic algorithm and backpropagation) neural networks. A soot-blowing optimization strategy was formulated by adopting the model of the maximum net heat profit and setting the accumulated system heat loss as the assessment variable. The applicability of the proposed monitoring model and optimization strategy was evaluated for the waterwall in a 650MWe coal-fired utility boiler. The monitoring results have verified that the change of system heat loss is in line with the actual slagging trend and the influence of the electric load change on the monitoring results is weakened greatly. The optimization results have shown that activating all soot blowers of the waterwall in every soot-blowing operation can achieve the higher net heat profit per unit time and the shorter duration for each pair of soot blowers. Using the optimized soot-blowing strategy can also realize the dynamic adjustment of the moment and the duration of soot-blowing, and improve the heat transfer performance of the waterwall remarkably.