BACKGROUND: Fast Ca2+ imaging using low-affinity fluorescent indicators allows tracking Ca2+ neuronal influx at high temporal resolution. In some systems, where the Ca2+-bound indicator is linear with Ca2+ entering the cell, the Ca2+ current has same kinetics of the fluorescence time derivative. In other systems, like cerebellar Purkinje neuron dendrites, the time derivative strategy fails since fluorescence kinetics is affected by Ca2+ binding proteins sequestering Ca2+ from the indicator. NEW METHOD: Our novel method estimates the kinetics of the Ca2+ current in cells where the time course of fluorescence is not linear with Ca2+ influx. The method is based on a two-buffer and two-indicator model, with three free parameters, where Ca2+ sequestration from the indicator is mimicked by Ca2+-binding to the slower buffer. We developed a semi-automatic protocol to optimise the free parameters and the kinetics of the input current to match the experimental fluorescence change with the simulated curve of the Ca2+-bound indicator. RESULTS: We show that the optimised input current is a good estimate of the real Ca2+ current by validating the method both using computer simulations and data from real neurons. We report the first estimates of Ca2+ currents associated with climbing fibre excitatory postsynaptic potentials in Purkinje neurons. COMPARISON WITH EXISTING METHODS: The present method extends the possibility of studying Ca2+ currents in systems where the existing time derivative approach fails. CONCLUSIONS: The information available from our technique allows investigating the physiological behaviour of Ca2+ channels under all possible conditions.