Written corrective feedback (WCF) is a crucial topic in the field of second language acquisition (SLA). The efficacy and importance of WRF has been questioned, even though there is evidence indicating that WCF favors learning and that it contributes cognitively to the improvement of accuracy levels in L2 (Loewen, 2012; Sheen, 2011). The objective of this article is to provide empirical evidence to SLA research and written L2 where no conclusive findings on the effects of WCF have yet been found, particularly on the effectiveness of metalinguistic corrective feedback in the learning of Spanish as a Foreign Language. To this end, a longitudinal experimental study was carried out to evaluate the short and long term effects of direct and indirect metalinguistic WCF on the writing process of SFL learners of levels A2 and B1. Focused errors treated with WCF correspond to those of written stress orthography, prepositions and grammatical concordance. The results indicate that indirect metalinguistic WCF is more effective than direct WCF, and that it supports significantly SFL learners' written accuracy in the short and long term. These findings allow us to obtain greater clarity on the delimitation of the relationship between error types, linguistic levels, written corrective feedback in SFL and metalinguistic accuracy.