This paper investigates the problem of global consensus between a complex dynamical network (CDN) and a known goal signal by designing an impulsive consensus control scheme. The dynamical network is complex with respect to the uncertainties, nonidentical nodes, and coupling time-delays. The goal signal can be a measurable vector function or a solution of a dynamical system. By utilizing the Lyapunov function and Lyapunov-Krasovskii functional methods, robust global exponential stability criteria are derived for the error system, under which global exponential impulsive consensus is achieved for the CDN. These criteria are expressed in terms of linear matrix inequalities (LMIs) and algebraic inequalities. Thus, the impulsive controller can be easily designed by solving the derived inequalities. Meanwhile, the estimations of the consensus rate for global exponential consensus are also obtained. Two examples with numerical simulations are worked out for illustration.