By "perfect" I'm going to assume you mean, "perfectly well-adapted to its environment". Clearly, then, the definition of "perfect" is environment-dependent.
Under this definition of perfect, then if we make a few assumptions, then the answer is "yes, that's true - kind of". Let me explain. Assume a fitness landscape that is unchanging with respect to time. In this case, evolution tends to push a species in the direction of "better adapted", and this will tend to lead a species to a "local maximum" in the landscape. This is "perfect" in the sense that small deviations from the local maximum are selected against. So to your statement is true. Actually, it's only kind of true - because there may be other maximums which are higher than the local maximum which evolution attains.
So we just considered the case where it is assumed that the fitness landscape doesn't change shape as time progresses. However, very few situations in biology adhere to this assumption. In the more general case where the fitness landscape is allowed to vary with time, evolution often goes 'round and 'round in circles, never attaining perfection. An example of this would be the evolution of human beauty. It will never get to the point where simply "everyone is beautiful" because that's not how sexual selection works. For more information, look up "sexual selection" and "evolutionary game theory". So therefore, evolution does not (in general) lead to the attainment of perfection.