This paper analyses immigrants' education-occupation mismatch as well as its impact on their wages in Spain. Using cross-sectional data from the National Immigrant Survey of Spain 2007, we estimate a probit model taking into account the possible problem of selection bias. We show that the incidence of immigrants' education-occupation mismatch in the Spanish labour market can largely be explained by the incidence of education-occupation mismatch in the last job held in the home country. The probability of having been over-educated at home shows to have a higher effect on the probability of being over-educated in the first job upon arrival where work experience gained in the home country shows to be highly valued by Spanish employers. In addition, our results show that those who were over-educated in their first job after arrival are more likely to continue in being overeducated in their current/last job in Spain. Furthermore, we analyse the performance of immigrants in Spain by estimating the wage penalty of over-education. Using log wage equation as well as predicted and counterfactual values distinguishing between immigrants being in the correctly matched occupation and those who are over-educated, we show that over-educated immigrants earn significantly lower wage compared to their correctly-matched counterparts, while over-educated immigrants' would have earned an even larger amount if the same individuals were employed in a correctly matched job instead. Significant differences are also apparent when restricting the models to the level of education.