This quasi-experimental study examined the effects of teacher scaffolding instruction on EFL students’ writing achievement and its consistency of the writing sub-skills and the proficiency-level groups. The study also explored the participants’ perceptions of teacher scaffolding instruction in their writing improvements. The participants were Grade 10 students, and data were collected using pre-post tests and close-ended questionnaires, and they all were analyzed quantitatively. The result of the independent samples test demonstrated that there was no statistically significant difference between the writing achievement mean score of the comparison group and the experimental group in the pretest (p=.955), showing that the study participants had comparable linguistic resources before commencing the study. Findings from a one-way MANOVA further proved that there was no statistically significant difference between the two groups pretest scores writing achievement in the combined dependent variables of using relevant content, vocabulary, grammatical structure, mechanic use, and organization as indicated by F (5, 90) =.615, p=0.689; Wilks Lambda=.967; partial eta squared (η2) =.033. The results proved students’ comparable linguistic backgrounds in their writing achievement at the outset. However, results from the paired-sample t-test showed a statistically significant difference between the pretest and post-test scores for the experimental group (p=.001), but insignificant differences for the comparison group (p=.426). An independent-sample t-test of the post-test also yielded a statistically significant difference between the two groups, favoring the experimental class (p=.001). The findings verified the teacher scaffolding instruction’s feasibility for improving students’ writing achievement. Results from a one-way MANOVA also proved that the writing subskills of content, vocabulary, grammatical structures, mechanics use, and organization jointly account for significant variance between the comparison and the experimental groups’ writing proficiency, in favor of the experimental group. Separate results of the variables revealed composing relevant content as the strongest predictor (F (1, 94) =35.08, p=.001), but no mechanics use differences between the two groups, F (1, 94) =.42, p=.520. The findings verified the instruction’s substantial role in enhancing learners’ overall writing achievement and all writing sub-skills except for mechanics use. Results from a one-way ANCOVA in pre-post-test scores unveiled the efficacy of teacher scaffolding instruction in improving EFL students’ writing skills for each proficiency level (p=0.001). After adjusting for the effect of the covariates, Bonferroni’s Pair-wise comparison at an α level of 0.05 further confirmed the low proficiency level group’s most significant writing improvement in the post-test compared to medium and high proficiency level groups. This implied that the independent variable (forms of teaching) influenced the dependent variable (writing achievement scores) significantly but differently between levels/groups, with the low proficiency level being the strongest predictor. Results from students’ questionnaire further affirmed participants’ positive perceptions of the efficacy of teacher scaffolding instruction for their writing improvements. Based on the findings, it is reasonable to conclude that teacher scaffolding instruction has promising effects on improving students’ writing skills.