The paper presents the novel meta-heuristic, called turbulent flow of water-based optimization (TFWO), algorithm to determine the optimal distribution of steel member sizes allocated to the truss structure that can safely sustain the specified design forces. The problem states the minimization of the cost function, described by the total weight of the designed structure, complying with the limit state specifications. The TFWO method performs the random searches among various whirlpool sets, where the best particle position of each group is pulled down by the centripetal (traction) force to the cavity in the center of a whirlpool. A centrifugal force acting in an opposite direction to the centripetal force randomly transfers the particle to the new position. The interaction between different whirlpools applies the individual centripetal forces to iteratively unify those surrounding whirlpools into ones with stronger tractions, and subsequently converges the optimal design solution. The accuracy of the TFWO scheme is illustrated through comparisons with some benchmarks processed by various recent optimization algorithms. These examples present the robustness of the proposed approach in the optimal design of steel structures at modest computing resources.