Effective semi-supervised learning (SSL) in medical im-age analysis (MIA) must address two challenges: 1) workeffectively on both multi-class (e.g., lesion classification)and multi-label (e.g., multiple-disease diagnosis) problems,and 2) handle imbalanced learning (because of the highvariance in disease prevalence). One strategy to explorein SSL MIA is based on the pseudo labelling strategy, butit has a few shortcomings. Pseudo-labelling has in generallower accuracy than consistency learning, it is not specifi-cally design for both multi-class and multi-label problems,and it can be challenged by imbalanced learning. In this paper, unlike traditional methods that select confident pseudo label by threshold, we propose a new SSL algorithm, called anti-curriculum pseudo-labelling (ACPL), which introduces novel techniques to select informative unlabelled samples, improving training balance and allowing the model to work for both multi-label and multi-class problems, and to estimate pseudo labels by an accurate ensemble of classifiers(improving pseudo label accuracy). We run extensive experiments to evaluate ACPL on two public medical image classification benchmarks: Chest X-Ray14 for thorax disease multi-label classification and ISIC2018 for skin lesion multi-class classification. Our method outperforms previous SOTA SSL methods on both datasets.