During escaped wildfires, community assets are at risk of damage or destruction. Preventive operations requiring dispatching resources and cooperation can be taken to protect these assets. The planning of such operations is sensitive to unforeseen disruptions that may occur. To account for the effects of the disruption, it may be necessary to alter the initial routes of the vehicles. The problem rising from the rescheduling of the vehicles is a bi-objective optimization problem known as the Dynamic Asset Protection Problem (D-APP). We propose a genetic algorithm based on the Non-dominated Sorting Genetic Algorithm (NSGA-II) to solve the D-APP. We define new mutation and crossover operators adapted to our problem, and we propose procedures to repair and evaluate a solution based on Mixed Integer Programming (MIP).