A new method called vectorized structure approach (VSA) is proposed for truss layout optimization, which allows effective changes to the truss structure by iteratively adding or removing nodes. This approach generalizes topology optimization to shape optimization, which is based on geometric construction rules. The VSA can generate structures containing a basic set of the required nodes and other nodes that may be critical for finding the optimal structure. The Covariance Matrix Adaptation Evolution Strategy (CMA-ES) is employed for VSA-based truss optimization. Numerical results show that the VSA-based CMA-ES is successful in finding optimal truss layouts and some of the found solutions are better than the best known topologies reported thus far in the literature.