Introduction: Untargeted metabolomics is a powerful tool to detect hundreds of metabolites within a given tissue and to compare the metabolite composition of samples in a comprehensive manner. However, with regard to pollen research such comprehensive metabolomics approaches are yet not well developed. To enable isolation of pollen that is tightly enclosed within the anthers of the flower, such as immature pollen, the current pollen isolation protocols require the use of a watery solution. These protocols raise a number of concerns for their suitability in metabolomics analyses, in view of possible metabolic activities in the pollen and contamination with anther metabolites. Objectives: We assessed the effect of different sample preparation procedures currently used for pollen isolation for their suitability to perform metabolomics of tomato pollen. Methods: Pollen were isolated using different methods and the metabolic profiles were analysed by liquid chromatography–mass spectrometry (LC–MS). Results: Our results demonstrated that pollen isolation in a watery solution led to (i) rehydration of the pollen grains, inducing marked metabolic changes in flavonoids, phenylpropanoids and amino acids and thus resulting in a metabolite profile that did not reflect the one of mature dry pollen, (ii) hydrolysis of sucrose into glucose and fructose during subsequent metabolite extraction, unless the isolated and rehydrated pollen were lyophilized prior to extraction, and (iii) contamination with anther-specific metabolites, such as alkaloids, thus compromising the metabolic purity of the pollen fraction. Conclusion: We conclude that the current practices used to isolate pollen are suboptimal for metabolomics analyses and provide recommendations on how to improve the pollen isolation protocol, in order to obtain the most reliable metabolic profile from pollen tissue.