In this research, we introduce a novel class of hybrid nonlinear conjugate gradient algorithms that generate an adequate search direction for descending gradients at each iteration. This property persists whether or not a line search is accurate and regardless of whether or not the objective function is convex. For functions that are not convex in general, we show that the suggested approaches converge globally under certain conditions. Numerical results demonstrate the effectiveness of these new hybrid methods in their application to selected testing problems compared to the standard Mevlana CG algorithm.