The graph of a fitness function for all possible solutions to a problem. Fitness landscapes are commonly used to describe the solution spaces of problems faced by evolutionary systems, including artificial evolutionary algorithms. The fitness function is graphed on the y-axis while all of the other parameters that make up the solutions are each graphed on their own axis.
The more inflection points that a fitness landscape has, the rougher it looks and the "harder" its problem is. This means that the probabilty that an evolutionary algorithm could reach the optimal solution of the problem faster than by trying every single solution and calculating which is the most fit would be smaller.