A new diagnostic test for regression models has been developed to address issues with existing tests. By using a residual marked empirical process based on projections, this test can accurately assess the goodness-of-fit of parametric regression models without relying on subjective parameters. The test is able to detect local alternatives converging to the null at the parametric rate and is robust to higher order dependence. Additionally, a new minimum distance estimator has been proposed, which inherits the good properties of the test. The consistency and asymptotic normality of the new estimator have been established, and Monte Carlo evidence supports the effectiveness of the testing procedure in econometric regression modeling.