Using cumulative residual processes, we propose joint goodness-of-fit tests for conditional means and variances functions in the context of nonlinear time series with martingale difference innovations. The main challenge comes from the fact the cumulative residual process no longer admits, under the null hypothesis, a distribution-free limit. To obtain a practical solution one either transforms the process in order to achieve a distribution-free limit or approximates the non-distribution free limit using a numerical or a re-sampling technique. Here the three solutions will be considered.It is shown that the proposed tests have nontrivial power against a class of root-n local alternatives, and are suitable when the conditioning information set is infinite-dimensional, which allows including models like autoregressive conditional heteroscedastic stochastic models with dependent innovations. The approach presented assumes only certain conditions on the first- and second-order conditional moments, without imposing any autoregression model. The test procedures introduced are compared with each other and with other competitors in terms of their power using a simulation study and a real data application. These simulations have shown that the statistical powers of tests based on re-sampling or numerical approximation of the original statistics are in general slightly better than those based on a martingale transformation of the original process.