In this paper, a generalized difference-based estimator is introduced for the vector parameter $eta$ in partially linear model when the errors are correlated. A generalized difference-based almost unbiased Liu estimator is defined for the vector parameter $eta$. Under the linear stochastic constraint $r=Reta+epsilon$, we introduce a new generalized difference-based weighted mixed almost unbiased Liu estimator. The efficiency properties of the difference-based weighted mixed regression method is analyzed. Finally, the performance of the new estimator is illustrated by a simulation study.